getpact

getpact / pact

Public

Authorization for AI agents

getpact.dev
14
0
100% credibility
Found May 11, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Pact is a secure authentication system for AI agents to read and write data using managed workspaces, permission policies, and tamper-proof audit logs.

How It Works

1
👋 Discover Pact

You learn about Pact, a simple way to let AI helpers safely access your team's data and tools without worries.

2
🏠 Create your space

Sign up with your email to make a secure workspace for your team and AI friends.

3
🔐 Set your rules

Choose who can do what, like deciding which friends can open which doors in your house.

4
🧠 Link your tools

Connect everyday services like chat apps or smart thinkers so your AI can use them.

5
🤖 Give AI access

Hand out special passes so your AI helpers can reach the tools they need.

6
📊 Watch it work

Your AI agents now chat, read, and act safely through protected paths.

✅ All secure

Everything runs smoothly with full records of what happened, so you stay in control.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is pact?

Pact delivers authorization for AI agents, letting them securely read/write data via JWTs issued from Google OIDC logins. Built in TypeScript on Cloudflare Workers with Postgres RLS, it includes a policy engine, tamper-evident audit chains, a forwarding gateway for upstream APIs (injecting GitHub authorization bearer tokens), and an MCP server exposing adapters like Slack as agent tools. Developers get CLI commands for workspace setup, token refresh, and audit verification.

Why is it gaining traction?

It stands out by baking in agents authorization at the edge—verifying policies and auditing every action without slow DB roundtrips—while avoiding vendor lock-in unlike cloud IAM services. The gateway handles GitHub authentication flows and bearer token injection securely, with SSRF guards and rate limits. MCP integration means agents in Claude or Cursor can call verified tools instantly, plus signed audit exports for compliance.

Who should use this?

AI engineers building agentic apps that access GitHub APIs, Slack channels, or custom "brains" under strict policies. Teams needing authorization for AI agents with full audit trails, especially those deploying on Cloudflare for low-latency verification. Ops folks managing multi-tenant workspaces via the admin API or pact CLI.

Verdict

Grab it for prototypes or internal AI tools—solid primitives for authorization github api flows and pact contract testing alternatives—but skip for high-scale prod until post-1.0. With 14 stars and 1.0% credibility, it's raw; expect rough edges in docs and adapters, but the audit/MCP combo hooks fast.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.