blockrunai/blockrun-mcp
blockrun-mcp is an open-source Model Context Protocol (MCP) server that provides AI assistants with unified, pay-per-use access to over 30 cutting-edge AI models from providers like OpenAI (GPT-5), Anthropic (Claude), Google (Gemini), xAI (Grok), and DeepSeek. Instead of managing multiple API keys, users connect a single USDC wallet on the Base blockchain to handle micropayments via the x402 protocol, automatically routing each request through the BlockRun gateway for billing and execution. The server acts as a proxy that can be integrated into any MCP-compatible client (such as Claude Code), enabling on-demand use of chat completions, image generation, video generation (e.g., Seedance), and search tools without per-provider credentials. Key features include a dynamic model tiering system (balanced, cheap, best) that automatically selects appropriate models based on task requirements, ensuring optimal cost-performance trade-offs. The server exposes a suite of tools beyond plain chat: it supports web research with Exa, prediction market queries, and video creation via Seedance models, all accessible through a consistent MCP interface. Pricing is transparent and per-request, with costs estimated before execution, and users can top up their wallet in USDC to pay only for what they use. The installation is straightforward, typically via npm, and configuration requires minimal setup—just a wallet address. The server runs locally, handling secure communication with the BlockRun gateway, and includes built-in caching and model fallback chains for reliability. It is designed for developers who want to experiment with or deploy multiple AI models in their workflows without the overhead of managing subscriptions or complex billing. With over 130 commits, 14 tagged releases, and a growing community (471 stars), blockrun-mcp is actively maintained, with frequent updates to add new models, optimize routing, and calibrate pricing. It represents a novel approach to AI model consum
- Category
- Developer Tools
- Pricing
- Open Source
Problem solved
Developers and AI enthusiasts face friction when trying to use multiple state-of-the-art language and generative models due to the need to sign up for, manage, and pay for separate API accounts across providers. blockrun-mcp eliminates this hassle by acting as a single gateway that requires only a USDC-funded wallet, allowing users to access over 30 models on a pay-per-use basis without individual API keys, thus lowering the barrier to experimentation and integration in development environments.
Use cases
- Augmenting an AI coding assistant like Claude Code with GPT-5 for advanced reasoning or Gemini for multilingual support without managing multiple API credentials.
- Generating AI videos using Seedance models directly from a development workflow, paying only for the seconds of video produced.
- Performing structured web research and content curation by combining AI-powered search tools with language models for summarization.
- Running prediction market queries via MCP to incorporate real-time forecasting data into automated decision-making pipelines.
- Quickly prototyping AI applications that compare outputs from different models to select the best result for a given task.
- Enabling hobbyist developers to experiment with state-of-the-art models like Grok or DeepSeek without upfront subscription costs.
Audience
- Developers
- ML Engineers
- Data Scientists
- Hobbyists
Pros
- Eliminates the need for multiple API keys by centralizing access through a single USDC wallet.
- Pay-per-use pricing with x402 micropayments keeps costs predictable and avoids unused subscriptions.
- Supports over 30 top-tier AI models, including GPT-5, Claude, Gemini, Grok, and DeepSeek.
- Seamless integration with MCP clients like Claude Code, with rich toolset for chat, search, and media generation.
- Open source and actively maintained, with regular model tier updates and pricing calibrations.
Cons
- Requires setting up a USDC wallet and funding it, which may be a barrier for non-crypto users.
- Pay-per-use costs can accumulate quickly for high-volume usage compared to flat-rate subscriptions.
- Dependence on the BlockRun gateway for routing, meaning availability and model selection are subject to their support and pricing changes.
Official website: https://github.com/blockrunai/blockrun-mcp
Canonical LambdaBase page: https://www.lambdabase.com/tools/blockrunai-blockrun-mcp