Quickstart

The fastest way to use Foundation Machines is to install the Sebastion AI GitHub App on a repo and let it review your next pull request. API access is available as a second step for power users and CI integrations.

Step 1, Install Sebastion AI on GitHub

1. Visit github.com/apps/sebastionai.
2. Click Install and choose the repositories you want covered. You can pick "All repositories" or a specific subset; you can change this later from your GitHub settings.
3. Push a pull request to one of those repositories.

Within about 60 seconds Sebastion opens a new GitHub issue on the same repository with structured findings, severity, file, line, message and suggested fix. No config files, no CI changes, no model picker.

The free tier uses Claude Sonnet 4.6 for audits. Paid tiers (Pro / Team / Enterprise) automatically route audits to Claude Opus 4.7. See models for the full per-plan map.

Step 2, API access (for power users)

Beyond the GitHub App, the same engine is exposed as an OpenAI-compatible API at https://api.foundationmachines.ai/v1. This is optional, and primarily useful for IDE integrations, CI scripts and custom automation.

Get an API key

Sign in at app.foundationmachines.ai with GitHub and mint a personal API key from the dashboard. Keys look like fm_<tenant>_<rand>.

Run a security audit on a snippet

curl -X POST https://api.foundationmachines.ai/v1/audit \
  -H "Authorization: Bearer fm_acme_..." \
  -H "Content-Type: application/json" \
  -d '{
    "snippets": [
      {
        "filename": "src/auth.ts",
        "language": "typescript",
        "content": "const AWS_SECRET = \"AKIAIOSFODNN7EXAMPLE\";\n"
      }
    ],
    "ruleset": "security"
  }'

You don't pass a model, Sebastion picks one based on your plan. See the audit reference for the full request and response schema.

Call the chat-completions gateway

You can also use the gateway as a drop-in replacement for OpenAI's chat endpoint. Here model is yours to choose, within the set allowed by your plan.

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.foundationmachines.ai/v1",
    api_key="fm_acme_...",
)

resp = client.chat.completions.create(
    model="claude-sonnet-4.6",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.foundationmachines.ai/v1",
  apiKey: process.env.FM_API_KEY!,
});

const resp = await client.chat.completions.create({
  model: "claude-sonnet-4.6",
  messages: [{ role: "user", content: "Hello" }],
});

curl

curl https://api.foundationmachines.ai/v1/chat/completions \
  -H "Authorization: Bearer fm_acme_..." \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-sonnet-4.6","messages":[{"role":"user","content":"Hello"}]}'

Check your usage

curl https://api.foundationmachines.ai/v1/usage/me \
  -H "Authorization: Bearer fm_acme_..."

Returns plan, daily token cap and what you've used today.