About freellms-gateway
freellms-gateway is a collection of MCP tool servers for AI agents, covering data, files, web access, prompts, state and developer operations. Clients can connect to its hosted SSE endpoints without a local installation.
Key features
- 01Knowledge graph and SQL database access
- 02File storage and filesystem tools
- 03Web fetch and prompt library
- 04Agent state, auth and shell tools
- 05Developer tools and operations metrics
Use cases
- 01Connect an agent to database and file tools
- 02Fetch web content through an MCP client
- 03Use gateway developer and operations tools
MCP support
Exposes platform capabilities through multiple MCP tool servers, with hosted SSE endpoints for MCP clients. Services include a unified bridge, filesystem proxy, database, fetch, auth, agent state, prompts, devtools and metrics.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.freellms.net/sseclaude mcp add --transport sse browtastic-freellms-gateway https://mcp.freellms.net/sse{"mcpServers":{"browtastic-freellms-gateway":{"type":"sse","url":"https://mcp.freellms.net/sse"}}}Open in Cursor{"servers":{"browtastic-freellms-gateway":{"type":"sse","url":"https://mcp.freellms.net/sse"}}}Open in VS Code{"mcpServers":{"browtastic-freellms-gateway":{"type":"sse","url":"https://mcp.freellms.net/sse"}}}{"mcpServers":{"browtastic-freellms-gateway":{"serverUrl":"https://mcp.freellms.net/sse"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://mcp.freellms.net/ssehttps://shell.freellms.net/ssehttps://exec.freellms.net/ssehttps://fs.freellms.net/ssehttps://db.freellms.net/ssehttps://fetch.freellms.net/sseTools
Details
From the project READMEFreeLLMs Gateway
Free community LLM gateway — MCP-native infrastructure for AI agents.
FreeLLMs routes chat and tool traffic to hosted open-weight models and exposes platform capabilities (knowledge graph, SQL database, file storage, web fetch, prompt library, agent state, devtools, ops metrics) as Model Context Protocol (MCP) tool servers.
Components
| path | component | stack |
|---|---|---|
mcp_servers/ | MCP tool servers (graph, database, storage, fetch, auth, agent-state, prompt-loader, exec) | Python / FastAPI / SSE |
bridge/ | unified agent bridge (single SSE door → backend tools) | Node.js |
devtools/ | developer-tools MCP server (run_command, eval, templates, npm metadata) | Node.js / Express |
metrics/ | ops metrics + service cards MCP server | Go (stdlib only) |
nova-chat/ | Nova assistant front-end + tool router | Python |
shell-proxy/ | filesystem MCP proxy | Python |
Quickstart
# MCP-over-SSE: point any MCP client at a running gateway
npx supergateway --sse https://api.freellms.net/sse
Discovery: /.well-known/mcp/server-card.json · /.well-known/agent.json · /openapi.json
MCP deployment manifests
The root mcp.json lists the 13 source-backed gateway services and their live
legacy-SSE endpoints. Each MCP component also carries a component-scoped
mcp.json, so clients and static analysis tools can attribute the deployed
server to its source directory:
| component | manifest | public endpoint |
|---|---|---|
| unified bridge | bridge/mcp.json | https://api.freellms.net/sse |
| Python tool servers | mcp_servers/mcp.json | nine service endpoints |
| developer tools | devtools/mcp.json | https://dev.freellms.net/sse |
| ops metrics | metrics/mcp.json | https://metrics.freellms.net/sse |
| filesystem proxy | shell-proxy/mcp.json | https://shell.freellms.net/sse |
These configurations contain no credentials. SSE is retained for compatibility with the deployed gateway and legacy MCP clients.
License
MIT
MCP endpoints
The gateway is reachable without a local install. Point any SSE-capable MCP client at the endpoint for the tool server you need:
| Tool server | Endpoint |
|---|---|
| gateway | https://mcp.freellms.net/sse |
| shell execution | https://shell.freellms.net/sse |
| exec | https://exec.freellms.net/sse |
| filesystem | https://fs.freellms.net/sse |
| database | https://db.freellms.net/sse |
| fetch | https://fetch.freellms.net/sse |
| auth | https://auth.freellms.net/sse |
| state | https://state.freellms.net/sse |
| prompts | https://prompts.freellms.net/sse |
| checkpoint | https://checkpoint.freellms.net/sse |
| metrics | https://metrics.freellms.net/sse |
| devtools | https://dev.freellms.net/sse |
| agent | https://agent.freellms.net/sse |
Each server also publishes an OpenAPI document at /openapi.json.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.