Skip to content
MCP server

Facet

Remote MCP server for building and calling small task-specific models from an agent

Responds, sign-in requiredStatus and history

The server answered 21 hours ago but needs an account or API key, so we could not list its tools.

Pricing
Freemium
Platforms
API, CLI
Runs
Remotely
Access
Needs an account or API key
Version
1.5.3
Listing updated
3 days ago

About Facet

Facet is a freemium MCP server for ai models and llm tooling. It runs as a hosted remote server, so there is nothing to install locally.

Facet turns one repeated LLM call in your codebase into a small dedicated model. Your coding agent describes the task and a few sample inputs, reviews a draft that lists the schemas, evals and training-row sources, and only then asks you to approve a quote for the build.

The docs say the agent sends nothing until you have seen it: rows come from tests, fixtures, sample files and few-shot examples rather than production logs, credential files are never opened, and rows that look like keys are dropped. Outputs are sent only when a person wrote them, not when a frontier model did.

Once a version is merged, the Model is called through the OpenAI SDK. The Claude Code plugin adds four commands: find, new, wire and status. Access is invite-only for now, via a waitlist.

Install

Run in a terminal

claude mcp add --transport http facet https://facet.halfspacelabs.com/mcp

Add to Settings → Connectors → Add custom connector

https://facet.halfspacelabs.com/mcp
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "facet": {
      "url": "https://facet.halfspacelabs.com/mcp"
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "facet": {
      "type": "http",
      "url": "https://facet.halfspacelabs.com/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "facet": {
      "serverUrl": "https://facet.halfspacelabs.com/mcp"
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "facet": {
      "type": "http",
      "url": "https://facet.halfspacelabs.com/mcp"
    }
  }
}

Tools this server provides (3)

Taken from the project documentation; the running server may differ.

create_model
Draft a Model from a task description and sample inputs, returning a review
start_change
Build v1 after the person approves the quote
request_change
Ask for a change to an existing Model

What it does

Draft before build
create_model returns a reviewable draft with schemas, evals and a quote before anything is made.
Approval-gated changes
A change starts only after a person approves its quote; API keys never approve a quote or merge.
Repo scan command
/facet:find ranks the LLM calls in your repo by what a Model could save, and sends nothing to Facet.
OpenAI SDK calls
A finished Model is called through the OpenAI SDK, so wiring it in is a client swap.
Several client options
Connects from Claude Code, Claude, Cursor, VS Code, ChatGPT and Codex, with OAuth sign-in.

Good for

  • Replace a costly, repeated LLM call with a small model
  • Rank which calls in a repo are worth replacing
  • Request an improvement to a live Model from your coding agent
  • Check how a Model scores on its evals before swapping it in

Pricing

Freemium, from $99 a month

Free tier with 5M serving tokens a month; Pay as you go; Team $99/month

Prices change. Check facet.halfspacelabs.com before you buy.

Facet FAQ

Do I need an API key?

Most clients sign in with an OAuth popup. A key is optional for clients that cannot sign in, but a key never approves a quote or a merge.

Is access open?

Not yet. Facet is invite-only and you join a waitlist first, then sign in with the invited email.

What data leaves my machine?

Only the description, row count, source files and a preview of rows, after you have seen and approved them. Production logs are not used unless you point at a file.

What does it cost?

Drafting is free. The first v1 trains free under a cap, then usage is billed; a Team plan is $99 a month. Each build is quoted first and the quote is the most it can cost.

Is Facet free?

Facet has a free plan, with paid plans for more usage or features. Paid plans start at $99 a month.

How do I add Facet to Claude Code?

Run this command in your terminal: claude mcp add --transport http facet https://facet.halfspacelabs.com/mcp

What are the best alternatives to Facet?

The closest alternatives to Facet are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

Alternatives to Facet

  1. REA

    MCP server

    CLI and MCP server that lets coding agents reverse engineer apps and binaries

    Runs locally
    • Open source
    • 464K GitHub stars
    • Developer tools
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. GitHub

    MCP server

    Official MCP server connecting AI assistants to GitHub repos, issues and workflows

    Not checked yet
    • Remote
    • Open source
    • 34.5K GitHub stars
    • Developer tools
  4. MCP toolkit giving coding agents symbol-level code retrieval and editing

    Runs locally
    • Open source
    • 31.5K GitHub stars
    • Developer tools
  5. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  6. Ouroboros

    MCP server

    Spec-first layer that pins requirements and verifies results of AI coding agent work

    Runs locally
    • Open source
    • 6.4K GitHub stars
    • Developer tools