Clelia (Astra) Bertelli 679d0bf876 Merge pull request #2 from run-llama/clelia/contribs-and-license
docs: contributing and license
2025-11-18 20:05:46 +01:00
2025-11-18 19:55:08 +01:00
2025-11-18 19:59:05 +01:00
2025-11-18 19:55:08 +01:00
2025-11-18 20:04:38 +01:00
2025-11-18 20:04:38 +01:00
2025-11-18 19:55:08 +01:00
2025-11-18 19:55:08 +01:00
2025-11-18 19:55:08 +01:00
2025-11-18 19:55:08 +01:00

PR Manager

Review Your PRs with Agentic AI

Are you looking for yet-another-agentic-companion to automate your GitHub PRs flow? Well, look no further: pr-manager, powered by LlamaIndex and Gemini 3 is all you need:

  • CLI native, with a beautiful and seamless TUI
  • Async- and streaming-first
  • Directly integrated with your personal GitHub thanks to an easy GitHub Application setup

Installation

You can install pr-manager with pip:

pip install pr-manager

Or clone this repository...

git clone https://github.com/run-llama/pr-manager
cd pr-manager/

... And install the package locally:

pip install .

Setup

In order to make the pr-manager work, you need to create a GitHub application with read/write permission on PRs (and install said application).

Once you did that, you need to generate a private RSA key, encode in base-64 and save it as GITHUB_PRIVATE_KEY in your .env file.

Here is an example script with which you can base64-encode your private key:

import base64

with open("private_key.pem", "rb") as f:
    content = f.read()

encoded = base64.b64encode(content).decode("utf-8")
print(encoded) # save the output as `GITHUB_PRIVATE_KEY` in your .env file

After that, save the app ID as the GITHUB_APP_ID env variable in your .env file.

In order for the agent to run, you need to have one last environment variable, i.e. GOOGLE_API_KEY.

Run

Once you completed the setup and everything is in place, you can start the agent with:

pr-manager start

And you can display its system prompt with:

pr-manager system

License and Contributions

This project is distributed under the MIT License.

Contributions are welcome and encouraged, and should follow the guidelines in CONTRIBUTING.md.

S
Description
Review your PRs with Agentic AI (powered by @run-llama and Gemini 3)
Readme MIT 235 KiB
Latest
2025-11-18 14:01:22 -05:00
Languages
Python 97.8%
Makefile 2.2%