onthemarkdata

onthemarkdata / petri

Public

Petri: Grow your own AI context.

18
1
100% credibility
Found Apr 15, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Petri orchestrates AI agents to decompose claims into verifiable facts, validate them through research and debate, and build structured knowledge graphs with citations.

How It Works

1
🔍 Discover Petri

You hear about Petri, a tool that helps AI deeply research and validate big ideas by breaking them into smaller truths.

2
🚀 Set up your research space

Create a simple folder for your project and prepare it with one command so everything is ready to go.

3
🌱 Plant your first idea

Type in a claim or question you want to explore, and watch AI break it down into a map of key supporting facts.

4
🤖 Grow the knowledge tree

Hit go, and teams of AI helpers research, debate, and check each fact step by step, building a trustworthy web of info.

5
📊 See your results live

Open the dashboard to browse the growing map, read summaries, check sources, and see what's solid or needs more work.

6
➕ Add new clues anytime

Drop in articles or links, and the system flags what to revisit, keeping your knowledge always fresh.

✅ Own verified insights

Enjoy a clean library of proven facts, connections, and sources your AI can use confidently forever.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 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 petri?

Petri is a Python CLI tool and dashboard for growing reliable AI context from claims, like culturing bacteria in a growth petri dish. Feed it a statement via `petri seed`, and it decomposes it into a DAG of sub-claims, then validates bottom-up through a Claude-powered pipeline: Socratic questioning, research with citations, critique debates, red-teaming, and evaluation. You get a searchable repo of verified facts, URLs, and relations—perfect for pre-loading agent knowledge without hallucinations.

Why is it gaining traction?

It stands out with agent-first UX: `petri grow` runs concurrent validations with live spinners, `petri feed` ingests new evidence and propagates changes automatically, and `petri launch` spins up a FastAPI dashboard for DAG viz and logs. Unlike basic RAG or petri net github tools, its event-sourced design tracks every agent step immutably, making audits trivial. Early buzz around petri github anthropic integration hooks devs tired of unverified LLM outputs.

Who should use this?

AI engineers building production agents needing grounded context, like research bots or decision systems. Research teams validating hypotheses (e.g., "open models catch frontier in 6 months") or devs curating petri dish growth medium for domain-specific knowledge. Skip if you're not ready for Claude API costs—thousands of tokens per colony.

Verdict

Alpha-stage (18 stars, 1.0% credibility) with solid docs and PyPI packaging, but monitor costs and watch for open-model support. Worth a spin for AI context builders; fork if it fits your petri growth workflow.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.