Gpt Chat
MCP server giving AI clients chat, vision, embedding and model-list tools for the OpenAI
- Pricing
- Open source
- Platforms
- CLI, Linux, macOS
- Version
- 0.1.0
- Listing updated
- today
Similar MCP servers
About Gpt Chat
Gpt Chat is an open-source MCP server for ai models and llm tooling. It is available on CLI, Linux, macOS, Windows and self-hosted.
The README says this Rust server wraps the OpenAI ChatGPT API and exposes it to MCP-compatible clients as four tools. Requests travel over stdio using JSON-RPC 2.0, so it plugs into desktop assistants that launch local MCP servers. Setup means supplying your own OpenAI API key in a TOML config file and building the binary with Cargo.
The chat tool accepts a prompt along with optional multi-turn message history, a system prompt, temperature, token limits and a JSON schema for structured output. Separate tools handle image analysis from a URL and text embeddings, while list_models returns the available model IDs. Default models are gpt-4o for chat and vision and text-embedding-3-small for embeddings.
It suits developers who already have OpenAI access and want their MCP client to call ChatGPT models directly. Usage is billed against the user's own OpenAI account, since the server itself does not add a hosted plan.
Tools this server provides (4)
Taken from the project documentation; the running server may differ.
- chat
- Send a chat completion request to ChatGPT with optional multi-turn history, system prompt, structured output and model selection
- chat_with_vision
- Analyse an image given an image URL and text prompt
- embedding
- Generate text embeddings using OpenAI's embedding model
- list_models
- List all available OpenAI models and their IDs
What it does
- Chat completions
- Sends prompts to ChatGPT with optional history, system prompt, temperature and token limits.
- Structured output
- Accepts a JSON schema string to constrain chat responses to a defined format.
- Vision analysis
- Analyses an image at a given http or https URL using a text prompt and a detail level setting.
- Text embeddings
- Generates embeddings for one string or an array of strings using an OpenAI embedding model.
- Model listing
- Returns the OpenAI model IDs available to the configured API key.
- Stdio transport
- Runs over stdio with JSON-RPC 2.0 so local MCP clients can launch it directly.
Good for
- Calling ChatGPT models from an MCP-compatible desktop assistant
- Asking GPT models to describe or analyse images from URLs
- Generating embeddings for search or retrieval pipelines inside an agent workflow
- Getting schema-constrained JSON responses from chat completions
- Checking which OpenAI models are available before choosing one
Pricing
Open source
Free and MIT-licensed; OpenAI API usage is billed separately to the user's account.
Gpt Chat FAQ
What do I need to run it?
The README lists Rust (edition 2024) and an OpenAI API key, which goes in ~/.config/mcp-server-gpt-chat/config.toml.
Which default models does it use?
Chat and vision default to gpt-4o, and embeddings default to text-embedding-3-small. Each can be overridden per request.
How is it connected to a client?
The README shows adding the built binary's path as a command entry in the Claude Desktop MCP configuration.
What license is it under?
The README states the MIT license.
Is Gpt Chat free?
Yes. Gpt Chat is open source, so you can use and self-host it for free.
What are the best alternatives to Gpt Chat?
The closest alternatives to Gpt Chat are Grok Chat, MCP Openai, MCP Aichat, Claude Chat and Gemma AI MCP.
Where this comes from
- Description: written from the project's README and its MCP Registry entry.