Gemini FAF

This Gemini FAF MCP Server provides a live HTTP endpoint that lets you retrieve Foundational AI-context DNA from .faf files for Gemini workflows. It enables you to integrate project DNA into AI agents, validate DNA, and reuse the same source of truth across your multi-agent environment.

How to use

You use this MCP server by querying the live HTTP endpoint to obtain the project DNA for a given FAF file. The endpoint acts as a Context Broker, returning the project DNA so your client can consume it in your agent workflows. You can fetch the DNA remotely or adapt a local workflow by parsing your own FAF files with supported client utilities.

Basic remote usage pattern involves sending a request to the endpoint with the path to your FAF file. The server responds with a structured payload containing the DNA and related readiness scores that you can feed into your agent’s decision-making process.

# Test the endpoint remotely for a project.faf
curl -X POST https://faf-source-of-truth-631316210911.us-east1.run.app \
  -H "Content-Type: application/json" \
  -d '{"path": "project.faf"}'

How to install

Prerequisites: you need Python and a package manager to install the MCP library.

Install the package from PyPI.

pip install gemini-faf-mcp

Use the included client helpers to interact with the live endpoint or parse local FAF files.

from gemini_faf_mcp import FAFClient, parse_faf, validate_faf

# Remote: call the live endpoint
client = FAFClient()
dna = client.get_project_dna()
print(f"Score: {client.get_score()}%")

# Local: parse your FAF file directly
data = parse_faf("project.faf")
result = validate_faf(data)
print(f"Tier: {result['tier']}")

Additional topics

Security and deployment details are provided for setup and integration in multi-agent environments. Follow recommended workflows for secret management and access controls when deploying to production.

The server exposes a live, scalable endpoint suitable for integration with various AI agents. You can use the endpoint as a single source of truth for project DNA, and you can validate and reuse the DNA in downstream workflows.

Available tools

FAFClient

Client interface to access the FAF DNA remotely via the live endpoint.

parse_faf

Parse a local FAF file to extract DNA for validation and scoring.

validate_faf

Validate FAF data and derive readiness tiers from the DNA.

get_project_dna

Retrieve the project DNA from the live FAF source of truth endpoint.