free-ai-mcp
Provides Google Gemini as an AI provider for the coding agent, enabling AI-powered code generation, file editing, and automated fixes, including vision support for image-based tasks.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@free-ai-mcpadd a name column to users"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
free-ai-mcp
Free AI coding agent that works like Claude Code — read, edit, search, fix, and commit using free AI APIs with automatic fallback. When one provider hits its rate limit, the next one takes over.
Works as both a CLI tool (interactive terminal agent) and an MCP server (use inside Claude Code, Cursor, etc.).
Features
Streaming responses — Tokens appear as the AI generates them, no waiting
Search/Replace editing — Precise SEARCH/REPLACE blocks (like Claude Code)
Auto-fix loop — Runs tests, reads errors, fixes code automatically (up to 3 attempts)
Multi-file edits — "add a name column to users" plans and edits migration + model + views
Project memory —
.free-ai.mdgives the AI persistent context about your projectGrep/search — Search across your codebase without leaving the agent
Smart commit — AI writes commit messages, you just confirm
Image/screenshot support — Send screenshots to Gemini vision, ask "build this UI"
Tab completion — Press Tab to autocomplete file paths
Conversation persistence — Chat history saved, resumes next session
Auto-retry — Bad AI response? Automatically retries with the next provider
Usage tracking — See request counts per provider with
/statusShell & Git — Run any command with
!prefix, git commands work directlyUndo — Revert any edit instantly
Auto-fallback — 5 free AI providers with automatic failover
Related MCP server: AI Collaboration MCP Server
Supported Providers (All Free)
Provider | Model | Free Tier | Sign Up |
Google Gemini | gemini-2.5-flash | 15 req/min, 1M tokens/day, vision | |
Groq | llama-3.3-70b | 30 req/min, 14,400 req/day | |
Mistral AI | mistral-small | Free tier | |
Cerebras | qwen-3.8-27b | Free tier | |
OpenRouter | nex-n2.5-pro (free) | Free models available |
Setup
1. Clone and install
git clone https://github.com/tmahajan90/free-ai-mcp.git
cd free-ai-mcp
npm install2. Get your free API keys
Sign up at any of the provider links above. You need at least one, but more = more fallback.
Recommended:
Google Gemini — best free tier (1M tokens/day) + image support
Groq — fastest responses (<1 second)
3. Configure API keys
cp .env.example .envEdit .env and paste your keys:
GEMINI_API_KEY=your_gemini_key_here
GROQ_API_KEY=your_groq_key_here4. (Optional) Project memory
Create .free-ai.md in your project root:
This is a Rails 8 app using Tailwind CSS v4.
Multi-tenant with acts_as_tenant.
Use Stimulus controllers for JS.
Tests are in test/ directory, run with: bundle exec rails testThe AI will use this context for every interaction.
Usage — Interactive Mode (like Claude Code)
cd your-project
free-aiEdit files
edit app/models/user.rb add email validation
add pagination to app/controllers/products_controller.rb
remove the header from app/views/layouts/application.html.erbMulti-file edits
add a name column to users
add a search feature to productsAuto-fix (run tests -> fix errors -> repeat)
fix Auto-detects test command
fix bundle exec rails test Custom test command
fix npm test Works with any runnerSearch across files
grep validates app/models Find text in files
search "def create" Same as grep
find TODO Find all TODOsImage / Screenshot (Gemini vision)
image screenshot.png Describe what's in the image
image mockup.png build this UI AI analyzes image and generates codeRead & explore
read app/models/product.rb
explain app/controllers/sales_controller.rb
review app/views/sales/_form.html.erb
ls app/models
scan app/controllersShell & Git
!bundle exec rails test Run any shell command
git status Git commands directly
commit AI writes commit messageAll commands
Command | Description |
| Search across files |
| Auto-fix: run tests, fix errors, repeat |
| Smart commit with AI-generated message |
| Show session changes / git diff |
| Analyze image with Gemini vision |
| Undo last file edit |
| Scan project structure |
| List files |
| Show providers + usage stats |
| Show help |
| Clear conversation + saved history |
| Exit (history saved for next session) |
Keyboard shortcuts
Tab — Autocomplete file paths
Up/Down — Navigate input history
Ctrl+C — Cancel current operation
Usage — One-shot CLI
free-ai edit app/models/user.rb "add email validation"
free-ai explain app/controllers/sales_controller.rb
free-ai review app/views/sales/_form.html.erb
free-ai generate app/models/invoice.rb "Rails model with validations"
free-ai ask "how to add pagination in Rails"
free-ai statusUsage — MCP Server (Claude Code / Cursor)
claude mcp add free-ai node /full/path/to/free-ai-mcp/src/mcp-server.jsTool | Description |
edit_code | Edit files with SEARCH/REPLACE blocks |
read_code | Read a file with line numbers |
generate_code | Generate a new file from instructions |
explain_code | Get an AI explanation of a file |
review_code | Code review for bugs & security |
ask_ai | Ask a coding question |
run_command | Run a shell command |
ai_status | Check configured providers |
How the edit engine works
Uses SEARCH/REPLACE blocks (same approach as Claude Code):
<<<<<<< SEARCH
exact lines from the original file
=======
replacement lines (or empty to delete)
>>>>>>> REPLACERetry chain:
Ask AI for SEARCH/REPLACE blocks
If blocks don't match the file -> retry full-file mode
If AI gives garbage response -> auto-retry with next provider
Falls back to full-file replacement as last resort
Tips
Add all 5 provider keys to maximize free usage across the day
Gemini has the most generous free tier (1M tokens/day) and supports images
Groq is the fastest (responses in <1 second)
Create
.free-ai.mdfor project-specific context the AI always knowsUse
fixafter making changes to auto-run and fix testsUse
commitfor AI-generated commit messagesTab-complete file paths to type faster
Conversation persists across sessions — no need to re-explain context
/statusshows how many requests you've made per provider
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
No tool schema history has been recorded yet.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Shared memory for AI coding agents. Save once, reuse from Cursor, Claude Code, Codex.
Connect AI assistants to GitHub - manage repos, issues, PRs, and workflows through natural language.
- OolkinOAuthcom.oolkin
AI colleagues that keep your standards, your project and their reasoning between sessions
Adaptive plan/build/review cycles for AI coding assistants, persisted across sessions.
Related MCP Servers
- AlicenseAqualityBmaintenanceEnables AI coding assistants to interact with OpenAI's Codex AI through the official CLI. Provides direct integration for code analysis, file review, and batch processing with zero API costs.3114MIT
- AlicenseNot gradedqualityNot gradedmaintenanceEnables collaboration with multiple AI providers (Claude, GPT-4, Gemini, Ollama) directly from VS Code with automatic project context injection and persistent conversation history. Provides streamlined tools for getting AI advice, multi-provider research, and enhanced context sharing across sessions.8-
- AlicenseNot gradedqualityDmaintenanceProvides AI-driven development tools including file system operations, multi-language code analysis with tree-sitter, Git operations, code execution, and system information retrieval.MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI-powered code review and improvement, including analysis, refactoring suggestions, and automatic test generation, with an optional agentic loop for iterative refinement.MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/tmahajan90/free-ai-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server