# GL1F Crypto: full text for language models Source: https://crypto.gl1f.com/ · generated 2026-10-02 · version 1.13.0 # GL1F Crypto: launch your own Crypto AI model > GL1F Crypto is a no-code crypto AI model builder on GenesisL1, an EVM Layer 1 blockchain (chain id 29). In the browser you build a dataset from Binance, Coinbase or Hyperliquid candles, train a gradient-boosted tree model with walk-forward validation, backtest it with fees, slippage and leverage, publish it as a Crypto AI model NFT and run inference that anyone can verify on-chain. Early alpha, educational, not financial advice. ## Build a Crypto AI model in five steps 1. **Build a dataset:** Pick Binance, Coinbase or Hyperliquid, a coin, a candle size and a yes-or-no question; the studio turns completed candles into signals and labels. 2. **Train the Crypto AI model:** Gradient-boosted trees learn from the past, walk-forward validation tests them on the future, and optional heuristic search tunes the settings. 3. **Backtest it:** Replay history with fees, slippage and stops to see the equity curve, drawdown and win rate. 4. **Deploy on-chain:** Mint the Crypto AI model as an Crypto AI Model NFT on GenesisL1; protocol fees are burnable by anyone. 5. **Run AI inference:** Score the latest candle or any past moment and verify the result on-chain. ## What a Crypto AI model answers A yes-or-no question about the next hours or days on one market, for example: "Will ETH rise 1% before it drops 0.5% within 5 hours?" or "Will DOGE drop 7% within the next 2 days?". It answers with a probability, computed by the same integer math in the browser and in the GenesisL1 EVM, so anyone can check a prediction on-chain. ## Key facts - Chain: GenesisL1, EVM Layer 1, chain id 29, live since 2021 - Model format: GL1F integer tree ensembles (gradient-boosted trees); the same integer score in the browser and inside the EVM - Data: Binance, Coinbase and Hyperliquid candles fetched by the browser; only completed candles, frozen to the exchange's server time - Signals: 274 in all: 212 from market data (momentum, volatility, volume, funding, BTC context) and 62 world signals (holidays and events, astronomy, macro and markets, disasters and conflict); list at https://crypto.gl1f.com/signals.txt - Validation: walk-forward folds (2 to 20), expanding or rolling windows, and a final test set kept locked until the end - Backtests: entries at the next candle's open, fees, slippage, stops, leverage from 1x to 100x, liquidation and account ruin - Crypto AI model NFTs: one per published model; its holder is the model admin; access is free, tips, a fee per inference, or subscriptions - Licenses: every model carries the license its creator chose at mint, from 38 standard licenses (SPDX License List 3.29.0); default CC BY-SA 4.0; paid models default to the GL1F On-Chain Use License (all rights reserved, use only through GenesisL1); a model's admin can only open its license up or move it to a later version - Terms of Service: version 1, stored in full on-chain; the registry owner publishes new versions and users accept them in the app - Fees: protocol fees (model creation, per-byte storage, listing) are burned; creator income goes to the model admin - Admin: the registry and marketplace owners set protocol fees, terms and the license catalog; ownership moves in two steps and can go to a multisig; nobody can withdraw protocol fees - Status: early alpha; educational; not financial advice; not for live or automated trading - Source: https://github.com/GenesisL1/gl1f-crypto (MIT) ## Questions and answers ### Which signals can a Crypto AI model use? 274 signals: 212 from the market itself (returns, trend, volatility, volume and taker flow, funding, BTC context) and 62 world signals: US and Asian holidays, Fed decision days, elections and expiries; Moon phases, Mercury, Venus and Mars retrogrades and planetary alignment; Treasury yields, the Fed funds rate, oil and natural gas, the dollar, VIX, the S&P 500, gold, inflation and unemployment; earthquakes, tropical storms and war attention. They load live in the browser, with years of history behind each one, and nothing to install. Every value uses only what was known when the candle closed. The full list is at signals.txt. ### What is a Crypto AI model? A machine-learning model that reads market signals such as recent returns, volatility, volume, funding and how Bitcoin is moving, and outputs the probability that your yes-or-no question comes true, for example: will ETH rise 1% before it drops 0.5% in the next 5 hours? GL1F Crypto builds it from exchange candles with gradient-boosted decision trees. ### Can AI predict crypto prices? Sometimes a little, often not at all. Markets are noisy, and most patterns are weak or vanish after fees. That is why every Crypto AI model here is tested walk-forward on data it never saw and backtested with costs, and why the honest answer can be: no edge. ### Is this an AI trading bot? No. GL1F Crypto is an educational Crypto AI model builder. It never places trades, never connects to exchange accounts or funds, and its Crypto AI models must not be used for live or automated trading. ### What makes it on-chain AI? The Crypto AI model is stored on GenesisL1 as an Crypto AI Model NFT, and its inference runs as integer math inside the EVM, so anyone can recompute a prediction on-chain and get exactly the score your browser shows. GenesisL1 EVM is your quant. ### What is GL1F Crypto? A free, open-source crypto AI studio that turns a market question into a machine-learning Crypto AI model. It downloads completed candles from Binance, Coinbase or Hyperliquid, trains gradient-boosted trees in your browser, backtests them with fees, and can mint the Crypto AI model as an Crypto AI Model NFT on GenesisL1 so anyone can verify its predictions. ### Do I need to code or know machine learning? No. Every field has a question mark with a plain-language explanation, and the defaults are sensible. You pick a coin, a target move, a stop and a time window; the studio does the rest. ### Will it make me money? Nobody can promise that, and we don't. GL1F Crypto is educational, experimental software. Most market patterns are weak and many disappear after fees. Crypto AI models must not be used for live trading. Crypto trading can lose you all of your money. ### Where does my data go? Nowhere. Candles are fetched straight from the exchange's public API into your browser, and training runs locally. Only deploying a model sends a transaction to GenesisL1. ### What happens to deployment fees? The model creation fee and the per-byte storage fee stay in the registry contract, and anyone can burn them to an unspendable address with one public call. No one, including the deployer, can withdraw them. ### Why does my Crypto AI model look great in training and bad in the backtest? That is usually overfitting, fees, or both, and it's exactly what the backtest is for. Try a simpler Crypto AI model (fewer trees, lower depth), turn on heuristic search, pick a bigger target or a longer candle, or accept that this question has no edge. ### Why does Hyperliquid start at 15-minute candles? Its public API keeps only the latest 5,000 candles per size. The signals need 22 days of warm-up history, which smaller candles can't cover, so the studio starts at 15m and sets the earliest possible start date for you. ### Can I use someone else's Crypto AI model? Yes. Published Crypto AI models can be run on the latest candle or any past moment and their scores verified on-chain; paid ones offer a fee per run or a subscription. Each shows the license it was published under. ### Which license does a Crypto AI model NFT have? The one its creator picked when minting it, from a catalog of standard licenses stored on GenesisL1: public domain (such as CC0), permissive (such as MIT, Apache 2.0, CC BY 4.0 or OpenMDW, made for machine-learning models), share-alike (such as CC BY-SA 4.0, GPL or ODbL) or restricted (non-commercial or no changes). The default is CC BY-SA 4.0. The license is written on-chain at mint; later its admin can only open it up or move it to a later version, even after a sale. ### Can someone copy a model from the chain? Technically yes: every model's bytes are public on GenesisL1. Legally it depends on the model's license. Open licenses allow copying; the GL1F On-Chain Use License, preselected for paid models, reserves all rights and allows use only through GenesisL1 on the admin's terms. The Terms of Service bind every user of the app and the contracts to respect each model's license. ## Links - [Studio](https://crypto.gl1f.com/app.html): build, backtest and publish a Crypto AI model ([Markdown](https://crypto.gl1f.com/app.md)) - [Marketplace](https://crypto.gl1f.com/market.html): published Crypto AI model NFTs ([Markdown](https://crypto.gl1f.com/market.md)) - [Docs](https://crypto.gl1f.com/docs.html) ([Markdown](https://crypto.gl1f.com/docs.md)) - [MCP server for AI agents](https://crypto.gl1f.com/api.html#mcp): https://crypto.gl1f.com/mcp (Streamable HTTP, no key), tools list_models, get_model, ask_model and trading_rules - [Yes threshold](https://crypto.gl1f.com/docs.html#threshold): a model's answer is yes when threshold <= P <= threshold_max (defaults 0.5 and 1) - [Terms of Service](https://crypto.gl1f.com/legal/terms.html) ([Markdown](https://crypto.gl1f.com/legal/terms.md)) - [GL1F On-Chain Use License](https://crypto.gl1f.com/legal/onchain-use-1.0.html) ([Markdown](https://crypto.gl1f.com/legal/onchain-use-1.0.md)) - [Everything in one file](https://crypto.gl1f.com/llms-full.txt) - [Source code](https://github.com/GenesisL1/gl1f-crypto) (MIT) --- # GL1F Crypto Studio: build a Crypto AI model without code > The studio runs entirely in your browser: market data comes straight from the exchange, training and backtests run on your device, and nothing is uploaded until you choose to publish. Open it at https://crypto.gl1f.com/app.html. Every step has help texts; the first visit asks you to accept the Terms of Service. ## The five steps ### 1. Build a dataset Pick Binance, Coinbase or Hyperliquid, a coin, a candle size and a yes-or-no question; the studio turns completed candles into signals and labels. ### 2. Train the Crypto AI model Gradient-boosted trees learn from the past, walk-forward validation tests them on the future, and optional heuristic search tunes the settings. ### 3. Backtest it Replay history with fees, slippage and stops to see the equity curve, drawdown and win rate. ### 4. Deploy on-chain Mint the Crypto AI model as an Crypto AI Model NFT on GenesisL1; protocol fees are burnable by anyone. ### 5. Run AI inference Score the latest candle or any past moment and verify the result on-chain. ## Saving and publishing - Save any trained model as a `.gl1f` file and load it again later. - Publishing stores the model on GenesisL1 and mints its Crypto AI model NFT. You choose its access (free, tips, a fee per inference, subscriptions) and its license. Paid models start with the GL1F On-Chain Use License, which reserves all rights; an open license on a paid model shows a warning. - Every published model has its own page, model.html?id=, drawn live from GenesisL1. ## Licenses in the studio The studio runs a published model on your device only when its license allows it. Models under the GL1F On-Chain Use License run in the browser only for their admin, active subscribers, or anyone while the model is free to run. ## Links - [Studio](https://crypto.gl1f.com/app.html): build, backtest and publish a Crypto AI model ([Markdown](https://crypto.gl1f.com/app.md)) - [Marketplace](https://crypto.gl1f.com/market.html): published Crypto AI model NFTs ([Markdown](https://crypto.gl1f.com/market.md)) - [Docs](https://crypto.gl1f.com/docs.html) ([Markdown](https://crypto.gl1f.com/docs.md)) - [MCP server for AI agents](https://crypto.gl1f.com/api.html#mcp): https://crypto.gl1f.com/mcp (Streamable HTTP, no key), tools list_models, get_model, ask_model and trading_rules - [Yes threshold](https://crypto.gl1f.com/docs.html#threshold): a model's answer is yes when threshold <= P <= threshold_max (defaults 0.5 and 1) - [Terms of Service](https://crypto.gl1f.com/legal/terms.html) ([Markdown](https://crypto.gl1f.com/legal/terms.md)) - [GL1F On-Chain Use License](https://crypto.gl1f.com/legal/onchain-use-1.0.html) ([Markdown](https://crypto.gl1f.com/legal/onchain-use-1.0.md)) - [Everything in one file](https://crypto.gl1f.com/llms-full.txt) - [Source code](https://github.com/GenesisL1/gl1f-crypto) (MIT) --- # GL1F Crypto Marketplace: Crypto AI model NFTs > Published Crypto AI models on GenesisL1, each one an NFT. Run them, subscribe to paid ones, buy and sell them. ## Access modes - **Free:** anyone runs the model. - **Tips:** anyone runs it; payments are optional. - **Paid:** a fee per inference, and optionally subscription plans that give an address access for a number of blocks. ## The model admin The holder of a model NFT is its admin: it sets access, fees, the fee recipient and subscription plans, can pause inference and can list the NFT for sale. A buyer becomes the new admin, and creator income follows the NFT: a fee recipient set by the previous owner stops applying after a sale or transfer. ## Licenses Every model shows the license it was published under, recorded on-chain. Its admin can move it only to a later version of the same license or to a more open license, never to a stricter one. ## Fees Creator income (fees per inference, tips, subscriptions) goes to the model admin. Protocol fees (creation, storage, listing) are burned; nobody can withdraw them. ## Links - [Studio](https://crypto.gl1f.com/app.html): build, backtest and publish a Crypto AI model ([Markdown](https://crypto.gl1f.com/app.md)) - [Marketplace](https://crypto.gl1f.com/market.html): published Crypto AI model NFTs ([Markdown](https://crypto.gl1f.com/market.md)) - [Docs](https://crypto.gl1f.com/docs.html) ([Markdown](https://crypto.gl1f.com/docs.md)) - [MCP server for AI agents](https://crypto.gl1f.com/api.html#mcp): https://crypto.gl1f.com/mcp (Streamable HTTP, no key), tools list_models, get_model, ask_model and trading_rules - [Yes threshold](https://crypto.gl1f.com/docs.html#threshold): a model's answer is yes when threshold <= P <= threshold_max (defaults 0.5 and 1) - [Terms of Service](https://crypto.gl1f.com/legal/terms.html) ([Markdown](https://crypto.gl1f.com/legal/terms.md)) - [GL1F On-Chain Use License](https://crypto.gl1f.com/legal/onchain-use-1.0.html) ([Markdown](https://crypto.gl1f.com/legal/onchain-use-1.0.md)) - [Everything in one file](https://crypto.gl1f.com/llms-full.txt) - [Source code](https://github.com/GenesisL1/gl1f-crypto) (MIT) --- # GL1F Crypto docs > GL1F Crypto is a no-code crypto AI model builder on GenesisL1, an EVM Layer 1 blockchain (chain id 29). In the browser you build a dataset from Binance, Coinbase or Hyperliquid candles, train a gradient-boosted tree model with walk-forward validation, backtest it with fees, slippage and leverage, publish it as a Crypto AI model NFT and run inference that anyone can verify on-chain. Early alpha, educational, not financial advice. [Overview](https://crypto.gl1f.com/docs.html#overview)[Contracts](https://crypto.gl1f.com/docs.html#contracts)[Fees & burns](https://crypto.gl1f.com/docs.html#fees)[Model monetization](https://crypto.gl1f.com/docs.html#monetization)[Licenses](https://crypto.gl1f.com/docs.html#licenses)[Admin & multisig](https://crypto.gl1f.com/docs.html#admin)[Live stats](https://crypto.gl1f.com/docs.html#stats) [Data & signals](https://crypto.gl1f.com/docs.html#data)[Validation](https://crypto.gl1f.com/docs.html#validation)[Backtests](https://crypto.gl1f.com/docs.html#backtests)[Verify it yourself](https://crypto.gl1f.com/docs.html#verify)[Web3 API](https://crypto.gl1f.com/docs.html#web3-api)[Yes threshold](https://crypto.gl1f.com/docs.html#threshold)[MCP server](https://crypto.gl1f.com/docs.html#mcp)[Track record](https://crypto.gl1f.com/docs.html#track-record)[Disclaimer](https://crypto.gl1f.com/docs.html#disclaimer)[Terms of Service](https://crypto.gl1f.com/legal/terms.html) **Docs** / Overview Contracts pending # How GL1F Crypto works. GL1F Crypto is a static web app. It downloads completed candles from Binance USD-M or Hyperliquid perpetuals, or Coinbase spot, straight into your browser, computes up to 212 causal market signals and 62 world signals, labels every candle with a target-before-stop question, trains integer gradient-boosted trees, and backtests them with fees. A Crypto AI model can be minted on GenesisL1 as a Crypto AI Model NFT; for a registered Crypto AI model, the local and on-chain evaluators return exactly the same integer score. Nothing runs on a server we control. There is no account, no API key, and no data leaves your device except the public transactions you sign. **01** / Contracts ## Dedicated crypto contracts, burnable fees. The crypto runtime gets its own contract set on GenesisL1 (chain id 29): a store for Crypto AI model bytes, a registry, a Crypto AI Model NFT, the on-chain runtime and a marketplace. The registry and marketplace are the GL1F contracts with a few changes: protocol fees are never paid out to anyone and accumulate until someone burns them; creators choose a license for each model; creator income follows the NFT when it changes hands; and contract ownership moves in two steps, so it can safely go to a multisig. **02** / Fees & burns ## Every protocol fee goes up in smoke. Deploying a Crypto AI model costs a flat **Crypto AI model creation fee** plus a **per-byte fee** for the bytes stored on-chain. Listing a Crypto AI Model NFT in the marketplace costs a **listing fee**. All three stay inside the contract that collected them. Anyone, at any time, can call `burnFees()` to send the whole balance to `0x000000000000000000000000000000000000dEaD`, an address no key controls. The owner can adjust fee levels; nobody can withdraw. Crypto AI model creation fee**—**set at deployment Per-byte fee**—**set at deployment Listing fee**—**set at deployment deploy fee = creation fee + per-byte fee × Crypto AI model bytes · paid in L1 · held by the registry until burned ### Burn what's waiting Burning is a public good: the caller pays only gas, and the event records who pressed the button. Registry · waiting to burn**—**creation + per-byte fees Marketplace · waiting to burn**—**listing fees Creator income is different: paid-inference fees, tips and access passes go to each Crypto AI model's admin (its current owner) or the recipient that admin chose, and are never burned. ## Model monetization and the marketplace. **The owner is the model admin.** Every Crypto AI model is an NFT. Whoever owns it is the model admin, the only address that can change its access mode, fee per inference, fee recipient and subscription plans, pause or resume inference, and list it for sale. When the NFT is sold or transferred, the new owner becomes the admin. ### Business models - **Free:** anyone can run the model and verify its scores on-chain. - **Tips:** inference stays open and users may pay the fee recipient voluntarily. - **Pay per inference:** each on-chain inference pays the fee the admin sets. - **Subscription plans:** for paid models the admin can publish plans (duration in blocks, price in L1, active or not). A subscriber pays once with `buyAccess` and their access key can run the model until the plan expires; buying again extends it. Durations are entered in days and converted at GenesisL1's measured block time. ### Marketplace An admin can list the Crypto AI Model NFT at a price (the marketplace takes a small listing fee, which is burnable like every protocol fee), cancel the listing, and anyone can buy it at the listed price. The sale price goes to the seller and the buyer receives the NFT and the admin rights. ### Who gets paid Inference fees, tips and subscription payments go to the fee recipient chosen by the current admin. When the NFT is sold or transferred, a recipient chosen by the previous owner stops applying and payments go to the new owner until it chooses its own. Sale prices go to the seller. Protocol fees (model creation, per-byte storage, listing) stay in the contracts until anyone burns them to 0x…dEaD. Nobody, including the contract owner, can withdraw them. The contract owner's powers are listed under [Admin & multisig](https://crypto.gl1f.com/docs.html#admin). ## One license per model, chosen at mint. Every Crypto AI model NFT carries a license for the model it holds. The creator picks it when minting, from a catalog of standard licenses stored in the registry; the global default, **CC BY-SA 4.0**, is preselected in the studio. The choice is written on-chain. Changing the default, closing an entry to new models or selling the NFT never changes it. Only the model admin can move it, and only to a later version of the same license or to a more open license, so rights already granted always stand. Anyone can read it with `licenseOf(tokenId)`, and model pages and the marketplace show it. Every model's bytes are public on GenesisL1, so no license can stop copying technically; the license decides what is lawful. For paid models the studio preselects the [GL1F On-Chain Use License](https://crypto.gl1f.com/legal/onchain-use-1.0.html): all rights reserved, use only through GenesisL1 on the admin's terms, and runs off-chain only for the admin, its subscribers, or anyone while the model is free to run. The [Terms of Service](https://crypto.gl1f.com/legal/terms.html) bind every user of the app and the contracts to respect each model's license, and the studio follows the same rules. The protocol admin can add licenses and new versions of a license to the catalog, close an entry to new models and change the default. It cannot change the license of a model that exists. Summaries are informal and are not legal advice; the linked license text is what counts. Identifiers follow the SPDX License List 3.29.0. OpenMDW 1.1 is newer than that list and uses LicenseRef-OpenMDW-1.1. ### Public domain No conditions at all. | # | License | SPDX | In short | |---|---|---|---| | 2 | [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/)default | `CC0-1.0` | Public domain dedication: anyone may use, change and sell it, with no credit required. | | 3 | [The Unlicense](https://unlicense.org/)default | `Unlicense` | Public domain dedication written for software: no conditions. | | 4 | [ODC PDDL 1.0](https://spdx.org/licenses/PDDL-1.0.html)default | `PDDL-1.0` | Open Data Commons public domain dedication for data and databases: no conditions. | | 5 | [BSD Zero Clause](https://opensource.org/license/0BSD)default | `0BSD` | Use for any purpose with no conditions, not even keeping the notice. | | 6 | [MIT No Attribution](https://github.com/aws/mit-0)default | `MIT-0` | The MIT license without the attribution condition: no conditions. | ### Permissive Use for anything, including commercially; keep the credit or notice the license asks for. | # | License | SPDX | In short | |---|---|---|---| | 7 | [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)default | `CC-BY-4.0` | Use, share and adapt for any purpose, including commercially; credit the creator and link the license. | | 8 | [MIT](https://opensource.org/license/MIT)default | `MIT` | Use for anything, including commercially; keep the copyright and license notice. | | 9 | [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0)default | `Apache-2.0` | Like MIT, plus an explicit patent grant; keep the notices and mark the files you change. | | 10 | [BSD 2-Clause](https://opensource.org/license/BSD-2-Clause)default | `BSD-2-Clause` | Use for anything; keep the copyright notice and the disclaimer. | | 11 | [BSD 3-Clause](https://opensource.org/license/BSD-3-Clause)default | `BSD-3-Clause` | BSD 2-Clause, plus: the authors' names may not be used to promote derived work. | | 12 | [ISC](https://www.isc.org/licenses/)default | `ISC` | Functionally the same as MIT, in simpler words. | | 13 | [zlib](https://opensource.org/license/Zlib)default | `Zlib` | Use for anything; do not misrepresent the origin, and mark altered versions as altered. | | 14 | [Boost 1.0](https://opensource.org/license/BSL-1.0)default | `BSL-1.0` | Use for anything; keep the license text with copies of the source. | | 15 | [UPL 1.0](https://opensource.org/license/UPL-1.0)default | `UPL-1.0` | Permissive, with an explicit patent grant; keep the license notice. | | 16 | [Blue Oak 1.0.0](https://blueoakcouncil.org/license/1.0.0)default | `BlueOak-1.0.0` | Modern permissive license with a patent grant; pass on the license text or a link to it. | | 17 | [Academic Free 3.0](https://opensource.org/license/AFL-3.0)default | `AFL-3.0` | Permissive, with a patent grant; keep the attribution notices. | | 18 | [OpenMDW 1.1](https://openmdw.ai/license/1-1/)default | `LicenseRef-OpenMDW-1.1` | Linux Foundation license made for machine-learning models: use for anything under copyright, patent, database and trade-secret rights; keep the license and notices. | | 19 | [OpenMDW 1.0](https://spdx.org/licenses/OpenMDW-1.0.html)default | `OpenMDW-1.0` | The first OpenMDW version for machine-learning models, listed by SPDX; same permissions as 1.1. | | 20 | [CDLA Permissive 2.0](https://cdla.dev/permissive-2-0)default | `CDLA-Permissive-2.0` | Linux Foundation data license: use and share freely; pass on the license text. Results you compute are unrestricted. | | 21 | [ODC Attribution 1.0](https://spdx.org/licenses/ODC-By-1.0.html)default | `ODC-By-1.0` | Open Data Commons attribution license for databases: use for anything; credit the source. | ### Share-alike (copyleft) Use for anything; versions you share must stay under the same license. | # | License | SPDX | In short | |---|---|---|---| | 1 | [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/)default | `CC-BY-SA-4.0` | Use for anything, including commercially; credit the creator, and share changed versions under the same license. | | 22 | [ODbL 1.0](https://opendatacommons.org/licenses/odbl/1-0/)default | `ODbL-1.0` | Database share-alike: credit the source; adapted databases you share publicly stay under ODbL. | | 23 | [CDLA Sharing 1.0](https://spdx.org/licenses/CDLA-Sharing-1.0.html)default | `CDLA-Sharing-1.0` | Linux Foundation data share-alike: data you share, including your additions, stays under the same terms. | | 24 | [MPL 2.0](https://www.mozilla.org/MPL/2.0/)default | `MPL-2.0` | File-level copyleft: changed MPL files stay MPL and their source is shared; may be combined with other work. | | 25 | [EPL 2.0](https://www.eclipse.org/legal/epl-2.0)default | `EPL-2.0` | Weak copyleft from the Eclipse Foundation: changes to EPL-licensed parts are shared under the EPL. | | 26 | [EUPL 1.2](https://spdx.org/licenses/EUPL-1.2.html)default | `EUPL-1.2` | European Union copyleft license, valid in all EU languages, with a list of compatible licenses. | | 27 | [OSL 3.0](https://opensource.org/license/OSL-3.0)default | `OSL-3.0` | Copyleft that also applies when the work is offered to others over a network. | | 28 | [LGPL 3.0 only](https://www.gnu.org/licenses/lgpl-3.0.html)default | `LGPL-3.0-only` | Weak copyleft, version 3 only: changes to the licensed work stay LGPL; other works may use it. | | 29 | [LGPL 3.0 or later](https://www.gnu.org/licenses/lgpl-3.0.html)default | `LGPL-3.0-or-later` | Weak copyleft, version 3 or any later version: changes to the licensed work stay LGPL; other works may use it. | | 30 | [GPL 3.0 only](https://www.gnu.org/licenses/gpl-3.0.html)default | `GPL-3.0-only` | Strong copyleft, version 3 only: anything distributed that is based on it must be GPL, with source. | | 31 | [GPL 3.0 or later](https://www.gnu.org/licenses/gpl-3.0.html)default | `GPL-3.0-or-later` | Strong copyleft, version 3 or any later version: anything distributed that is based on it must be GPL, with source. | | 32 | [AGPL 3.0 only](https://www.gnu.org/licenses/agpl-3.0.html)default | `AGPL-3.0-only` | GPL 3.0 only, plus: offering a changed version as a network service also requires sharing its source. | | 33 | [AGPL 3.0 or later](https://www.gnu.org/licenses/agpl-3.0.html)default | `AGPL-3.0-or-later` | GPL 3.0 or later, plus: offering a changed version as a network service also requires sharing its source. | ### Restricted (not open source) Standard Creative Commons licenses that forbid commercial use, changed versions, or both. | # | License | SPDX | In short | |---|---|---|---| | 34 | [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/)default | `CC-BY-NC-4.0` | Non-commercial use only; credit the creator. Changed versions allowed. | | 35 | [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/)default | `CC-BY-NC-SA-4.0` | Non-commercial use only; credit the creator; share changed versions under the same license. | | 36 | [CC BY-ND 4.0](https://creativecommons.org/licenses/by-nd/4.0/)default | `CC-BY-ND-4.0` | Share unchanged copies only, commercial use allowed; credit the creator. No changed versions may be shared. | | 37 | [CC BY-NC-ND 4.0](https://creativecommons.org/licenses/by-nc-nd/4.0/)default | `CC-BY-NC-ND-4.0` | Share unchanged copies for non-commercial purposes only; credit the creator. | ### Reserved (no copying) All rights reserved. The model may be used only through GenesisL1 on its admin's terms; copying it is not allowed, even though its bytes are public. | # | License | SPDX | In short | |---|---|---|---| | 38 | [GL1F On-Chain Use 1.0](https://crypto.gl1f.com/legal/onchain-use-1.0.html)default | `LicenseRef-GL1F-OnChain-Use-1.0` | All rights reserved: use it only through GenesisL1 on the terms its admin sets (free, per run or by subscription). No copying, sharing or derived models. Its admin and active subscribers may also run it themselves, and anyone may while it is free to run. | ## Who can change what: admins and multisigs. The [protocol admin page](https://crypto.gl1f.com/admin.html) does all of this from a wallet: it deploys the contracts, sets fees, burns, publishes Terms, manages licenses and hands ownership over in two steps; when the owner is a multisig it prepares a Safe Transaction Builder batch instead of sending. ### Model admin The owner of a Crypto AI model NFT runs that model: access, fees, fee recipient, subscription plans, pausing and sales. To put a model under a multisig, transfer the NFT to the multisig; its income follows. Owner-signed free inference needs a single-key signature, so a multisig that owns a paid model can give an ordinary wallet an access key with `setOwnerAccessKey`. ### Protocol admin The registry and the marketplace each have an owner. Together they can: - set the model creation fee and the per-byte fee (registry) and the listing fee (marketplace); - publish new versions of the [Terms of Service](https://crypto.gl1f.com/legal/terms.html), stored in full on-chain, add licenses and new versions of a license, close a license to new models and choose the default license; - cancel any marketplace listing; A model's admin can delete it once no paid subscription is still running, so nobody is owed access. Deleting burns the NFT and removes the model from the registry, the marketplace and search; the bytes already written to GenesisL1 storage stay on-chain, since deployed code cannot be erased. Until then the admin can pause inference. The protocol admins cannot delete models, cannot withdraw protocol fees, change a model's license, bytes or settings, move anyone's NFT, or point the registry at a different NFT contract (that wiring happens once, at deployment). Every change is a public transaction. ### Handing the protocol admin to a multisig 1. Set up the multisig on GenesisL1 (chain id 29) and check that it can send a transaction there. 2. The current owner calls `transferOwnership(multisig)` on the registry and on the marketplace. Nothing changes yet. 3. The multisig calls `acceptOwnership()` on each contract and takes control. Until then the old owner stays in charge and can cancel with `transferOwnership(0x0)`, so a mistyped address, or a multisig that only exists on another chain, can never lock the contracts. Terms of Service in force: version 1, as bundled with this site. Each new version is recorded in the registry with `setToS`; the app asks every user to accept it, and publishing a model requires it. `scripts/owner_tx.mjs` in the repository prints the transaction data for each step and for fee, license and terms changes, ready for a multisig's custom transaction form. **03** / Live stats ## Read straight from the chain. These numbers are view calls against GenesisL1, made by your browser when the page loads. Access mix Fees Latest Crypto AI models **04** / Data & signals ## Completed candles, causal signals. - **Binance USD-M** (`fapi.binance.com`): OHLCV, quote volume, trade count, taker buy volume, funding history. All 212 signals. - **Coinbase spot** (`api.exchange.coinbase.com`): OHLCV. Signals needing trade counts, taker flow or funding are unavailable, leaving 193. Native candles are 1m, 5m, 15m, 1h, 6h and 1d; other sizes are aggregated from the next smaller native size and labelled as derived. - **Hyperliquid perps** (`api.hyperliquid.xyz/info`): OHLCV, trade counts and hourly funding; no taker flow, leaving 204 signals. The API serves only the latest 5,000 candles per coin and size, so candles start at 15m and the studio sets the earliest start date for you (15m ≈ 52 days of history including the 22-day warm-up, 1h ≈ 208 days). Candles you have already downloaded stay in the browser cache. Memecoins use a k prefix (kPEPE); 6h candles are built from 2h. - Only completed candles are used, frozen to the exchange's server time. Coinbase and Hyperliquid publish nothing for buckets without trades; those gaps are filled with the previous close and zero volume, but a hole longer than 6 hours is treated as an outage and stops the build. Funding signals use each venue's own published rates (Binance every 8 hours, Hyperliquid hourly), so a Crypto AI model belongs to the venue it was trained on. - Every signal at candle *t* uses candles up to and including *t*. Values are cast to binary32 and quantized with GL1F's rule, clampint32(⌊Q·x + ½⌋), exactly as the trainers and the on-chain runtime do. Warm-up history is fetched before your start date so long indicators are complete from the first row. - Candles are cached in your browser (IndexedDB). Clearing site data removes them. ## Web3 API: run any model from code. The [Web3 API page](https://crypto.gl1f.com/api.html) has the full reference, the addresses and a bot example. Every Crypto AI model can be called from your own code, in a browser, in Deno or in Node 18+, with the GL1F Crypto helper [sdk/gl1f-crypto.js](https://crypto.gl1f.com/sdk/gl1f-crypto.js) and `ethers` v6. The Inference step's **Web3 API** panel shows ready code for the open model, its contract addresses and RPC. import * as ethers from "ethers"; import { GL1FCrypto } from "./gl1f-crypto.js"; const gl1f = new GL1FCrypto({ ethers, rpcUrl, chainId: 29, registry, runtime }); const model = await gl1f.model(42); // title, question, pricing, plans, inputs const engine = await GL1FCrypto.nodeEngine(); // browser: await GL1FCrypto.browserEngine() const inputs = await gl1f.latestInputs(model, { engine }); // latest completed candle, the studio's own engine const out = await gl1f.predict(model, inputs.valuesQ); // { probability, yes, answer, scoreQ, via } const now = await gl1f.ask(model, { engine, threshold: 0.6 }); // the same in one call, yes at 0.6 or more | Model | Call | What it costs | |---|---|---| | Free or tips | `predict(model, valuesQ)`: a read of `predictView` | Nothing: no key, no gas | | Paid, with a plan | `predict(model, valuesQ, { accessKey })`: the access key signs the request (EIP-712 `AccessView`) and `predictAccessView` checks it | The plan, bought once for the key's address with `buyAccess(model, planId, key.address, signer)`; each call is a free read | | Paid, per run | `predict(model, valuesQ, { payer })`: a `predictTx` transaction | The model's fee per run, paid to its admin, plus gas | | Your own paid model | `predict(model, valuesQ, { owner })`: your wallet signs (`OwnerView`) | Nothing | **Access keys.** `gl1f.newAccessKey()` makes a key pair; buy a plan for its address once and keep its private key on your server or bot. Signatures expire after 5 minutes by default and cover the exact inputs, so a key cannot be replayed for other inputs. `accessStatus(model, address)` tells until which block a key works. The studio can create a key and buy a plan for it in the Web3 API panel. Inputs are the model's signals on a completed candle, quantized to integers (`quantize(values, scaleQ)`). `latestInputs` computes them with the published market engine ([sdk/gl1f-engine.js](https://crypto.gl1f.com/sdk/gl1f-engine.js)), the code the studio runs, so the score equals the studio's and the chain's. Models that use world signals fetch the same public data. ## Yes or no: your threshold. A Crypto AI model gives a probability **P** that the answer to its question is yes, for the latest completed candle. By default the answer is **Yes** when P is 0.5 or more. You can set your own rule: a stricter **threshold** (Yes only at 0.6 or more, say) and an optional **upper limit**, so the answer is Yes only inside a range (from 0.6 to 0.85, say). The rule never changes the probability, only the yes or no read from it. | Where | How | |---|---| | Model page | In **Ask the model**: "Answer Yes when the probability is from … to …" (default 0.50 to 1.00). The answer on screen follows the fields at once, without running the model again, and a link can carry them: `model.html?id=1&threshold=0.6&threshold_max=0.85`. | | Studio inference | **Signal threshold** and **Upper limit**; the receipt records both. | | Backtest | **Entry threshold** and **Upper P limit**: a trade opens only when P is in the range. The threshold table compares lower thresholds under your upper limit. | | Web3 API | `predict(model, valuesQ, { threshold, thresholdMax })` and `ask(model, { engine, threshold, thresholdMax })` return `yes` and `answer`; `decide(probability, { threshold, thresholdMax })` applies the rule to a probability you already have. | | MCP server | `ask_model` and `trading_rules` take `threshold` and `threshold_max`. | Both limits are numbers from 0 to 1, the threshold not above the upper limit; the defaults are 0.5 and 1. A higher threshold usually leaves fewer signals: before trusting one, check in the backtest how many trades it leaves and whether nearby thresholds do about as well. Educational, not investment advice. ## MCP server: models for AI agents. AI assistants such as Claude and ChatGPT connect to outside tools through the Model Context Protocol (MCP). GL1F Crypto has an MCP server at `https://crypto.gl1f.com/mcp`: connect it, and the assistant can list Crypto AI models, read one, ask it for its current answer and get the rules its backtest trades by. It is read-only and needs no key, wallet or login. | Tool | What it does | |---|---| | `list_models` | Published models, newest first: number, title, question, market and candle, access (free, tips or paid) and page. | | `get_model` | One model's details, with the trading rules its backtest uses. | | `ask_model` | Its answer for the latest completed candle: yes or no, the probability, the candle and the model's page. Optional `threshold` and `threshold_max` (defaults 0.5 and 1). Free and tip models. | | `trading_rules` | The backtest's rules: direction, the signal range, entry, target and stop from the EMA baseline, and the time limit. | **Connect it.** Claude: add a custom connector with the address above. ChatGPT: turn on developer mode, then add a connector (app) with the address and no authentication. Claude Code: `claude mcp add --transport http gl1f-crypto https://crypto.gl1f.com/mcp`. Any MCP client that speaks Streamable HTTP works, and local agents can run it over stdio from the repository (`deno task stdio` in `mcp/`). It runs a model the same way as the model page, once per candle for everyone who asks: the inputs from public exchange data, the answer from GenesisL1. Paid models are not available through it, and each client may send 120 requests a minute. The code is in the repository's `mcp/` folder (Deno, MIT), with `deploy/setup-mcp.sh` to run your own. ## A track record anyone can check. When you publish a Crypto AI model, the studio can add its training report to the model's on-chain record: the final test window and row count, ROC AUC, log-loss, accuracy and a SHA-256 fingerprint of every test row (open time, label, integer score). The studio computes it; nobody can type it in. It rebuilds the same test rows from public candles with the backtest replay and publishes the report only if the result matches the training session exactly. After minting the report can never change. In the Inference step, anyone who can run the model (anyone for a free model; the admin and subscribers of a paid one) can press **Verify with on-chain inference**: the studio rebuilds the rows, compares every metric and the fingerprint with the published report, and checks sampled scores against GenesisL1's on-chain inference. It also scores every completed candle since the model was minted, data it could not have seen; the mint time is recorded on-chain. A report is the creator's own test, so the record since minting is the stronger evidence. **Private internals.** Any model can keep its number of trees, depth and signal list private on the marketplace, model pages, preview cards and in the studio, for everyone but its admin. It is on by default for paid models and the admin can change it at any time after minting. The bytes stay public on-chain: this makes copying harder, it is not encryption. ## World signals: calendars, the sky, macro, disasters. Besides 212 market signals, a Crypto AI model can learn from 62 world signals. Holidays, events and astronomy are computed from each candle's close time. Macro, disaster and war signals are fetched live in your browser when you build a dataset, straight from public sources that allow it, with no key and nothing to install. Every value uses only what was known when the candle closed. The plain name, machine name and meaning of all 274 signals are in [signals.txt](https://crypto.gl1f.com/signals.txt). | Group | Examples | Source and timing | |---|---|---| | Holidays & events | US market holidays, Fed decision days, US and major world elections, Lunar New Year, Golden Weeks, options expiry, US jobs day | Computed: NYSE holiday rules and built-in Fed, election and holiday lists | | Astronomy | Moon phase and illumination, full and new moon, Mercury, Venus and Mars retrograde, planetary alignment, days to equinox or solstice | Computed: JPL approximate orbital elements and a low-precision lunar theory | | Macro & markets | 10-year and 2-year Treasury yields, yield curve, Fed funds rate, WTI and Brent oil, natural gas, US dollar index, VIX, S&P 500, gold, CPI inflation, unemployment | Live: Federal Reserve and BLS data via DBnomics; EIA, BLS, Federal Reserve H.10, Cboe and Shiller data via DataHub. Daily series count from 2 days after their date, monthly ones from 40 to 45 days after the month | | Disasters & conflict | M6+ and M7+ earthquakes, active hurricanes and typhoons, their winds, major storms, new storms, war attention, days since a major war escalation | Live: USGS (since 1973), NASA EONET tropical storms (since 2000) and Wikipedia pageviews (since July 2015), counted up to the day before; escalations from a built-in list | If a live source cannot be reached, the studio uses the copy this site publishes at [data/world.json](https://crypto.gl1f.com/data/world.json), refreshed by its scheduled job. Every world signal has years of history: the signal list shows where each one starts, and a dataset that starts before a selected signal's history is stopped with a clear message. Fed decision dates, elections and war escalations are built-in lists from 2014 that need a yearly update. A model that uses world signals loads the same data again when it runs inference. **05** / Validation ## Tested on the future, never the past. The final temporal test partition (the last 15% of rows) is fixed before any development folds are built, and it is scored once. Walk-forward validation (the default) uses five expanding-window folds, starting from the first 40% of development rows. Every earlier/later boundary is protected by the label's outcome look-ahead (your horizon), followed by purge and embargo margins, which are zero by default; no training row's outcome window reaches the rows it is evaluated on. Each fold's early stopping picks the number of trees; candidates are compared by their mean validation loss across completed folds, the most recent fold supplies the retained tree budget, and the final Crypto AI model is fitted on that fold's training rows or, with refit, on all development rows before the final test is scored. Shuffled validation is available for experiments and is clearly labelled as optimistic. The number of folds is adjustable from 2 to 20, the window can be expanding or rolling, and the first-window and final-test shares can be changed in the studio. **Heuristic search (optional).** Instead of one training run, the studio can try up to 200 candidate settings. The first candidate is exactly yours; each later one varies trees, depth, learning rate, minimum leaf size and patience around the best candidate so far, the same moves as the GenesisL1 Forest studio. Every candidate is scored on all walk-forward folds, the lowest mean validation log-loss wins, and only then is the final test evaluated, once. A live chart can show train and validation loss tree by tree while this runs. Scores reported: ROC AUC, log-loss, balanced accuracy, precision and recall at 0.5, Brier score, the confusion matrix, and the baseline of always predicting the majority answer. **06** / Backtests ## The rules of the test drive. - The worker replays the Crypto AI model's exact signals over the chosen period with the same engine as the training dataset. The signals warm up on the 120 days before the period (or from the training data's start, when that is later), so any past period can be tested, including one before the training data. - A signal is a candle whose probability is at or above the entry threshold, and at or below the upper P limit when one is set (by default 1: no limit). The trade enters at the next candle's open, long for Up Crypto AI models and short for Down Crypto AI models. - Target and stop are the label's levels, measured from the EMA baseline at the signal candle. If both are touched in one candle, the stop wins. A candle that opens beyond a level exits at that open. Signals whose entry price is already past the target or the stop are skipped and counted separately. - Without a hit, the trade exits at the close of the last horizon candle. One position at a time; by default the whole account is the margin at 1× (no leverage), and both can be changed (see below). - Fees are charged on entry and exit; slippage worsens both fills. Funding payments, partial fills, liquidity and outages are not modelled. - The default period is the Crypto AI model's final test period, which it never trained on. Earlier periods are in-sample and flagged. **Entries and leverage.** A signal exists only after its candle has closed, so the earliest trade is the next candle's open. There is no trade if that candle had no trades (a gap-filled bucket), if there is no next candle, or if its open is already beyond the target or the stop the signal defined. One position at a time; when the target and the stop are both touched in one candle, the stop is assumed first. Leverage, any whole number from 1× to 100×, multiplies returns on margin and fees on the full notional; with isolated margin, a position is liquidated when the adverse move reaches about 1 / leverage minus the maintenance margin, losing that trade's margin. Funding payments are not included. Coinbase spot runs without leverage. **07** / Verify it yourself ## Don't trust. Re-run. Every contract address above is checked against the keccak-256 hash of its deployed bytecode before the app reads or writes anything. Published Crypto AI models are rebuilt from on-chain chunks and hashed against their Crypto AI model ID. On the Inference step, "Verify on GenesisL1" asks the on-chain runtime for the same prediction and compares the integer score. The source code, contracts and tests are public in the [GL1F Crypto repository](https://github.com/GenesisL1/gl1f-crypto), derived from [GenesisL1 Forest](https://github.com/GenesisL1/Forest). ## Disclaimer **GL1F Crypto is open-source, experimental data-science software provided for educational and research purposes only.** It is not a trading product, signal service, broker, exchange, fund or adviser. - **No investment advice.** Nothing in the software, its documentation, its example questions or any Crypto AI model is financial, investment, trading, legal or tax advice, or a recommendation or solicitation to buy, sell or hold any asset. - **Serious risk.** Trading crypto-assets, and especially derivatives, leveraged products and memecoins, carries a high risk of rapid and total loss. Only you can decide what risk is appropriate for you. - **Crypto AI models can be wrong.** Predictions are statistical estimates from historical data. They can be overfit, stale, biased by data errors or broken by changing market conditions. Past performance, including any backtest, does not indicate future results. - **Backtests are simulations.** They ignore many real-world effects such as liquidity, partial fills, latency, funding, outages and your own behaviour. - **Not for live setups.** Crypto AI models built or run with this software must not be used for live trading, trading bots, copy-trading or any automated execution. - **No warranty, no liability.** The software is provided "as is", without warranty of any kind, express or implied. In no event shall the authors, contributors or copyright holders be liable for any claim, damages or other liability arising from its use. - **Permanent transactions.** Deployments, listings and burns on GenesisL1 are public and irreversible. Protocol fees are burned and cannot be refunded. - **Your responsibility.** You are responsible for your decisions, for securing your wallet and keys, and for complying with the laws and regulations that apply to you. Binance, Coinbase and Hyperliquid are trademarks of their respective owners and are named only to identify public market-data sources. GL1F Crypto is not affiliated with, endorsed by or sponsored by any of them. L1 coin is referred to only in its operational role as the unit in which GenesisL1 fees are paid. ### No warranty of any kind The software, the website, the smart contracts, the Crypto AI models and all data are provided “as is” and “as available”, without warranty of any kind, express or implied, including but not limited to warranties of merchantability, fitness for a particular purpose, title, non-infringement, accuracy, reliability, security, availability, or that results will be error-free. The software may contain bugs, errors, defects, inaccuracies, omissions and other imperfections, and may behave differently from its description. The smart contracts have not been audited and may contain vulnerabilities. Market data comes from third parties and may be wrong, delayed or incomplete. To the maximum extent permitted by law, in no event shall the authors, contributors, Decentralized Science Labs, the GenesisL1 Association or any affiliated party be liable for any claim, loss or damage of any kind (direct, indirect, incidental, special, consequential or exemplary), including loss of funds, profits, data or opportunity, arising from or in connection with the software, its results or its use, whether in contract, tort or otherwise, even if advised of the possibility of such damage. You use it entirely at your own risk. --- # License catalog | # | License | SPDX | Group | In short | |---|---|---|---|---| | 1 | [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/) | CC-BY-SA-4.0 | share-alike | Use for anything, including commercially; credit the creator, and share changed versions under the same license. | | 2 | [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/) | CC0-1.0 | public-domain | Public domain dedication: anyone may use, change and sell it, with no credit required. | | 3 | [The Unlicense](https://unlicense.org/) | Unlicense | public-domain | Public domain dedication written for software: no conditions. | | 4 | [ODC PDDL 1.0](https://spdx.org/licenses/PDDL-1.0.html) | PDDL-1.0 | public-domain | Open Data Commons public domain dedication for data and databases: no conditions. | | 5 | [BSD Zero Clause](https://opensource.org/license/0BSD) | 0BSD | public-domain | Use for any purpose with no conditions, not even keeping the notice. | | 6 | [MIT No Attribution](https://github.com/aws/mit-0) | MIT-0 | public-domain | The MIT license without the attribution condition: no conditions. | | 7 | [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) | CC-BY-4.0 | permissive | Use, share and adapt for any purpose, including commercially; credit the creator and link the license. | | 8 | [MIT](https://opensource.org/license/MIT) | MIT | permissive | Use for anything, including commercially; keep the copyright and license notice. | | 9 | [Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0) | Apache-2.0 | permissive | Like MIT, plus an explicit patent grant; keep the notices and mark the files you change. | | 10 | [BSD 2-Clause](https://opensource.org/license/BSD-2-Clause) | BSD-2-Clause | permissive | Use for anything; keep the copyright notice and the disclaimer. | | 11 | [BSD 3-Clause](https://opensource.org/license/BSD-3-Clause) | BSD-3-Clause | permissive | BSD 2-Clause, plus: the authors' names may not be used to promote derived work. | | 12 | [ISC](https://www.isc.org/licenses/) | ISC | permissive | Functionally the same as MIT, in simpler words. | | 13 | [zlib](https://opensource.org/license/Zlib) | Zlib | permissive | Use for anything; do not misrepresent the origin, and mark altered versions as altered. | | 14 | [Boost 1.0](https://opensource.org/license/BSL-1.0) | BSL-1.0 | permissive | Use for anything; keep the license text with copies of the source. | | 15 | [UPL 1.0](https://opensource.org/license/UPL-1.0) | UPL-1.0 | permissive | Permissive, with an explicit patent grant; keep the license notice. | | 16 | [Blue Oak 1.0.0](https://blueoakcouncil.org/license/1.0.0) | BlueOak-1.0.0 | permissive | Modern permissive license with a patent grant; pass on the license text or a link to it. | | 17 | [Academic Free 3.0](https://opensource.org/license/AFL-3.0) | AFL-3.0 | permissive | Permissive, with a patent grant; keep the attribution notices. | | 18 | [OpenMDW 1.1](https://openmdw.ai/license/1-1/) | LicenseRef-OpenMDW-1.1 | permissive | Linux Foundation license made for machine-learning models: use for anything under copyright, patent, database and trade-secret rights; keep the license and notices. | | 19 | [OpenMDW 1.0](https://spdx.org/licenses/OpenMDW-1.0.html) | OpenMDW-1.0 | permissive | The first OpenMDW version for machine-learning models, listed by SPDX; same permissions as 1.1. | | 20 | [CDLA Permissive 2.0](https://cdla.dev/permissive-2-0) | CDLA-Permissive-2.0 | permissive | Linux Foundation data license: use and share freely; pass on the license text. Results you compute are unrestricted. | | 21 | [ODC Attribution 1.0](https://spdx.org/licenses/ODC-By-1.0.html) | ODC-By-1.0 | permissive | Open Data Commons attribution license for databases: use for anything; credit the source. | | 22 | [ODbL 1.0](https://opendatacommons.org/licenses/odbl/1-0/) | ODbL-1.0 | share-alike | Database share-alike: credit the source; adapted databases you share publicly stay under ODbL. | | 23 | [CDLA Sharing 1.0](https://spdx.org/licenses/CDLA-Sharing-1.0.html) | CDLA-Sharing-1.0 | share-alike | Linux Foundation data share-alike: data you share, including your additions, stays under the same terms. | | 24 | [MPL 2.0](https://www.mozilla.org/MPL/2.0/) | MPL-2.0 | share-alike | File-level copyleft: changed MPL files stay MPL and their source is shared; may be combined with other work. | | 25 | [EPL 2.0](https://www.eclipse.org/legal/epl-2.0) | EPL-2.0 | share-alike | Weak copyleft from the Eclipse Foundation: changes to EPL-licensed parts are shared under the EPL. | | 26 | [EUPL 1.2](https://spdx.org/licenses/EUPL-1.2.html) | EUPL-1.2 | share-alike | European Union copyleft license, valid in all EU languages, with a list of compatible licenses. | | 27 | [OSL 3.0](https://opensource.org/license/OSL-3.0) | OSL-3.0 | share-alike | Copyleft that also applies when the work is offered to others over a network. | | 28 | [LGPL 3.0 only](https://www.gnu.org/licenses/lgpl-3.0.html) | LGPL-3.0-only | share-alike | Weak copyleft, version 3 only: changes to the licensed work stay LGPL; other works may use it. | | 29 | [LGPL 3.0 or later](https://www.gnu.org/licenses/lgpl-3.0.html) | LGPL-3.0-or-later | share-alike | Weak copyleft, version 3 or any later version: changes to the licensed work stay LGPL; other works may use it. | | 30 | [GPL 3.0 only](https://www.gnu.org/licenses/gpl-3.0.html) | GPL-3.0-only | share-alike | Strong copyleft, version 3 only: anything distributed that is based on it must be GPL, with source. | | 31 | [GPL 3.0 or later](https://www.gnu.org/licenses/gpl-3.0.html) | GPL-3.0-or-later | share-alike | Strong copyleft, version 3 or any later version: anything distributed that is based on it must be GPL, with source. | | 32 | [AGPL 3.0 only](https://www.gnu.org/licenses/agpl-3.0.html) | AGPL-3.0-only | share-alike | GPL 3.0 only, plus: offering a changed version as a network service also requires sharing its source. | | 33 | [AGPL 3.0 or later](https://www.gnu.org/licenses/agpl-3.0.html) | AGPL-3.0-or-later | share-alike | GPL 3.0 or later, plus: offering a changed version as a network service also requires sharing its source. | | 34 | [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) | CC-BY-NC-4.0 | restricted | Non-commercial use only; credit the creator. Changed versions allowed. | | 35 | [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) | CC-BY-NC-SA-4.0 | restricted | Non-commercial use only; credit the creator; share changed versions under the same license. | | 36 | [CC BY-ND 4.0](https://creativecommons.org/licenses/by-nd/4.0/) | CC-BY-ND-4.0 | restricted | Share unchanged copies only, commercial use allowed; credit the creator. No changed versions may be shared. | | 37 | [CC BY-NC-ND 4.0](https://creativecommons.org/licenses/by-nc-nd/4.0/) | CC-BY-NC-ND-4.0 | restricted | Share unchanged copies for non-commercial purposes only; credit the creator. | | 38 | [GL1F On-Chain Use 1.0](https://crypto.gl1f.com/legal/onchain-use-1.0.html) | LicenseRef-GL1F-OnChain-Use-1.0 | reserved | All rights reserved: use it only through GenesisL1 on the terms its admin sets (free, per run or by subscription). No copying, sharing or derived models. Its admin and active subscribers may also run it themselves, and anyone may while it is free to run. | --- # GL1F Crypto Terms of Service Version 1 · Effective 4 October 2026 These terms are an agreement between you and Decentralized Science Labs LLC, a Wyoming limited liability company ("we", "us") about GL1F Crypto: the app and website at crypto.gl1f.com and any copy of them (the "App"), and the GL1F Crypto smart contracts on GenesisL1 (the "Contracts"), together the "Service". By using the App, or by interacting with the Contracts through any interface, you accept these terms. If you do not accept them, do not use the Service. The version in force is recorded on-chain in the GL1F Crypto registry (`tosVersion`, `tosText`). If a copy of these terms differs from the on-chain text, the on-chain text applies. ## 1. What the Service is 1. GL1F Crypto is open-source, experimental, educational software in early alpha. It may contain bugs, errors and imperfections, and it may change or stop at any time. 2. It is not a broker, exchange, trading venue, fund, investment adviser or signal service. Nothing in the Service is financial, investment, trading, legal or tax advice. 3. Crypto AI models, backtests and predictions are statistical estimates that can be wrong. They must not be used for live or automated trading. 4. The App runs in your browser. Market data comes from third parties (Binance, Coinbase, Hyperliquid) under their own terms. We never hold your funds, keys or data. ## 2. Your responsibilities 1. You must be of legal age and able to accept these terms where you live, and your use of the Service must be lawful, including under sanctions laws. 2. You are responsible for your wallet, keys, transactions, taxes and decisions. 3. Transactions on GenesisL1 are public, permanent and cannot be reversed. We cannot undo, refund or recover anything. ## 3. Crypto AI models and their licenses 1. Every published Crypto AI model is stored on GenesisL1, and anyone can technically read its bytes. What anyone may do with a model is decided by its license, recorded on-chain (`licenseOf(tokenId)`) and shown in the App. 2. **You must respect the license of every model**, whether you reach it through the App, the Contracts or any other way. For a model under the [GL1F On-Chain Use License](https://crypto.gl1f.com/legal/onchain-use-1.0.html) this means in particular: no copying, sharing or selling it, no models derived from it, and no use of a copy to avoid its fees or access terms. 3. When you publish a model, you confirm that you have the rights to do so and that it infringes no one's rights, and you grant everyone the license you select. The Contracts record that license for the model. 4. Only the model's admin can change its license later, and only to a later version of the same license or to a more open license. Rights already granted under an earlier license stay in force. 5. Predictions you obtain in line with a model's license are yours to use for any lawful purpose, subject to these terms. ## 4. Model NFTs, admins and income 1. The holder of a Model NFT is the model's admin. The admin sets access, fees and subscription plans, and can sell the NFT. 2. When a Model NFT is transferred, by sale or otherwise, the previous holder transfers to the new holder, to the extent the law allows, the rights it holds in the model as its licensor and admin. Licenses and access already granted continue. 3. Inference fees, tips and subscription payments go to the current admin or the recipient it chose. Sale prices go to the seller. Protocol fees are burned by the Contracts and are never refunded. ## 5. How the App respects licenses 1. The App shows every model's license, applies the access terms recorded on-chain, and never offers published models for download. 2. The App runs a model under the GL1F On-Chain Use License on your device only when that license allows it: for its admin, for holders of active access, or while the model is free to run. Otherwise it can be used only through the Contracts. 3. You may not use the App, its code or the Contracts to get around a license or access terms. ## 6. Acceptable use You may not use the Service to break the law, infringe anyone's rights, manipulate markets, publish malicious or deceptive content, present predictions as investment advice, attack or overload the App, the Contracts or GenesisL1, or help anyone do these things. ## 7. Smart contracts and protocol administration 1. The Contracts are unaudited and may contain vulnerabilities. 2. The owner of the registry and of the marketplace, which may be a multisig, can set protocol fees, publish new versions of these terms, add licenses and new license versions to the catalog, close licenses to new models and cancel marketplace listings. It cannot withdraw protocol fees, change a model's license or move anyone's NFT. The owner cannot delete models. A model's admin can delete their own model only when no paid subscription to it is still running; deletion burns the model's NFT and removes it from the registry, while the bytes already written to GenesisL1 storage remain on-chain. Every such action is a public transaction. ## 8. Changes to these terms We may publish new versions of these terms. A new version takes effect when the registry owner records it on-chain, and the App asks you to accept it before you continue. Publishing a model always requires accepting the version in force. If you do not accept a new version, stop using the Service. Earlier versions remain readable on-chain in past transactions. ## 9. Privacy and cookies The Site keeps a few settings in your browser so that it works: your theme, your cookie choice and that you accepted these terms. It uses statistics cookies only if you agree, and you can withdraw that at any time with Cookie settings at the bottom of every page. The app asks market, public-data and blockchain services for data directly from your browser. The [Cookie policy](./cookies.html) explains what is kept, for how long and why. ## 10. Open source The App's code is available under the MIT License. That license covers the code, not the Crypto AI models published with it, which are covered by their own licenses. ## 11. No warranty The Service, the App, the Contracts, the models, the data and all content are provided "as is" and "as available", without warranty of any kind, express or implied, including merchantability, fitness for a particular purpose, title, non-infringement, accuracy, reliability and availability. ## 12. Limitation of liability To the maximum extent the law allows, we and the authors, contributors and affiliated parties are not liable for any loss or damage of any kind arising from the Service, its results or its use, including loss of funds, profits or data, whether in contract, tort or otherwise. Where liability cannot be excluded, it is limited to the minimum the law allows. ## 13. Indemnity To the extent the law allows, you will compensate us for claims and costs that third parties bring against us because you broke these terms or a model's license. ## 14. Governing law These terms are governed by the laws of the State of Wyoming, United States, without regard to its conflict-of-law rules. The state and federal courts located in Wyoming have jurisdiction, unless the mandatory law of your country of residence gives you the right to go to court elsewhere. Nothing in these terms takes away the protection that the mandatory consumer law of your country of residence gives you. ## 15. Contact Questions and notices: open an issue at [github.com/GenesisL1/gl1f-crypto](https://github.com/GenesisL1/gl1f-crypto) or write to [GenesisL1 on X](https://x.com/genesis_L1). --- # GL1F On-Chain Use License Version 1.0 · SPDX identifier: `LicenseRef-GL1F-OnChain-Use-1.0` · Published by Decentralized Science Labs LLC (the "Steward") **In short (not part of the license):** the model is public on GenesisL1, but all rights are reserved. You may use it through GenesisL1 on the terms its admin sets: free, a fee per run, or a subscription. Its admin and active subscribers may also run it themselves, and so may anyone while it is free to run. You may not copy it, share it, sell it, build other models from it, or use a copy to avoid its fees. ## 1. Definitions - **Model:** the Crypto AI model recorded in the GL1F Crypto registry on GenesisL1 under this license, including its bytes, trees, parameters, structure and metadata. - **Model NFT:** the token that represents the Model in the GL1F Crypto registry. - **Licensor:** whoever holds the rights in the Model: the creator who published it and, after a transfer of the Model NFT, the holder of those rights at the time. - **Model Admin:** the current holder of the Model NFT. - **Runtime:** the GL1F Crypto smart contracts on GenesisL1 that evaluate the Model, including evaluations that GenesisL1 nodes answer without a transaction. - **Access Terms:** the access mode (free, tips or paid), fees and subscription plans that the Model Admin has set in the registry at the time of use. - **Access Holder:** an address that holds access to the Model through an active subscription or an owner access key recorded in the registry. - **Use:** evaluating the Model on inputs to obtain predictions. - **You:** anyone exercising rights under this license. ## 2. All rights reserved The Licensor reserves all rights in the Model except the permissions this license grants. Recording the Model on a public blockchain makes its bytes readable by anyone. It does not give anyone permission to copy, use, change or share the Model beyond this license. ## 3. Operating the network Anyone may store, replicate, transmit, cache, index, archive and display the Model's bytes as part of operating GenesisL1 and as part of services that read or present the chain as a whole, such as nodes, validators, RPC providers, archives, indexers and block explorers. This permission does not cover offering the Model on its own as a file, dataset, download or service for running it. ## 4. Using the Model through GenesisL1 You may Use the Model through the Runtime on the Access Terms, including by paying the fees and holding the access they require. You may use the predictions You obtain this way for any lawful purpose. ## 5. Running the Model yourself You may also Use the Model outside the Runtime, for Your own purposes and without sharing it, only: 1. while You are the Model Admin or an Access Holder; or 2. while the Access Terms let anyone run the Model without paying (free or tips mode). Software that does this for You, such as the GL1F Crypto app, may load the Model's bytes temporarily for that purpose. It may not keep them, show them or offer them for download beyond that purpose. ## 6. Restrictions Except as this license expressly allows, or as applicable law allows despite this license, You may not: 1. copy, download, extract or store the Model or its parameters, except temporary copies needed for a permitted use; 2. publish, distribute, sell, rent, lend, sublicense or otherwise make the Model, or any copy or part of it, available to anyone, including as a file, dataset, application, API or service; 3. use a copy of the Model to avoid its Access Terms, for example by running a copy instead of paying per run or subscribing; 4. change the Model, combine it with other models, or make derived models from it, including by retraining, fine-tuning or distillation, or by training on its predictions to imitate it; 5. remove, change or hide this license notice or the Model's identity on GenesisL1. ## 7. Model NFT transfers When the Model NFT is transferred, the new holder becomes the Model Admin, and the Licensor's rights follow the Model NFT as set out in the [GL1F Crypto Terms of Service](https://crypto.gl1f.com/legal/terms.html). Permissions You already received under this license, such as access already paid for, continue for their term. ## 8. Termination Your permissions under this license end automatically if You break it. If You stop the breach within 30 days after You learn of it, Your permissions are restored, once. Sections 2, 6, 9 and 10 survive termination. ## 9. No warranty The Model and its predictions are provided "as is", without warranty of any kind, express or implied, including fitness for a particular purpose, accuracy and non-infringement. Predictions are statistical estimates, not investment advice. ## 10. Limitation of liability To the extent the law allows, the Licensor and the Steward are not liable for any loss or damage arising from the Model, its predictions or this license. ## 11. Versions The Steward may publish new versions of this license under new version numbers. A Model is governed by the version recorded for it in the GL1F Crypto registry. Its Model Admin may move it to a later version of this license, or to a more open license, by recording that in the registry; the change applies from the block in which it is recorded and does not shorten access already paid for. A Model can never be moved to a more restrictive license. ## 12. Which models this license covers This license covers each Model whose license recorded in the GL1F Crypto registry is `LicenseRef-GL1F-OnChain-Use-1.0`. The registry function `licenseOf(tokenId)` identifies the Model, its license and the block from which that license applies.