Skip to content

Conversation

@1bcMax
Copy link

@1bcMax 1bcMax commented Jan 10, 2026

Summary

This PR adds BlockRun.AI as a new LLM provider, enabling access to multiple AI models (including Claude!) through x402 micropayments on Base chain.

Why BlockRun?

  • Claude access without API keys - Closes AI Model: Claude #257! Users can now use Claude Sonnet 4 and other Anthropic models via BlockRun
  • Decentralized LLM access - Addresses AI Model: Decentralized hosted LLM #322! Pay-per-request with crypto, no monthly subscriptions
  • Multi-model gateway - Single wallet integration for GPT-4o, Claude, DeepSeek, Gemini
  • x402 native - Uses the same payment protocol you already support in skills/x402!

Available Models via BlockRun

Model Description
openai/gpt-4o GPT-4o
openai/gpt-4o-mini GPT-4o Mini
anthropic/claude-sonnet-4 Claude Sonnet 4 ⭐
deepseek/deepseek-chat DeepSeek V3
google/gemini-2.0-flash Gemini 2.0 Flash

Security

Your private key never leaves your machine. BlockRun uses EIP-712 signatures for payment authorization - the same security model as MetaMask transactions.

Changes

  • Add BLOCKRUN to LLMProvider enum
  • Add BlockRunLLM class with LangChain-compatible wrapper
  • Add blockrun_wallet_key config option
  • Add BlockRun models to llm.csv
  • Update docs/llm.md with setup instructions
  • Add blockrun-llm dependency

Setup

# Set your Base chain wallet private key
export BLOCKRUN_WALLET_KEY=0x...

# Fund wallet with USDC on Base
# Use any BlockRun model!

Links


Closes #257
Addresses #322

Add BlockRun.AI as a new LLM provider using x402 micropayments on Base chain.
This enables access to multiple LLM providers (OpenAI, Anthropic, DeepSeek, Google)
with wallet-based payments - no API keys needed.

Changes:
- Add BLOCKRUN to LLMProvider enum
- Add BlockRunLLM class with LangChain-compatible wrapper
- Add blockrun_wallet_key config option
- Add BlockRun models to llm.csv (GPT-4o, Claude Sonnet 4, DeepSeek, Gemini)
- Update docs/llm.md with BlockRun setup instructions
- Add blockrun-llm dependency to pyproject.toml

Closes crestalnetwork#257 (Claude support)
Addresses crestalnetwork#322 (Decentralized hosted LLM)
@hyacinthus
Copy link
Collaborator

Thank you for your code contribution. Since using x402 to pay for LLM API is a new concept, we will need some time to check for conflicts with existing logic, so the review will be a bit slower.

@1bcMax
Copy link
Author

1bcMax commented Jan 12, 2026

@hyacinthus Thank you for your reply. Please let me know what I can help to bring clarity.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

AI Model: Claude

2 participants