dev-agents 0.10.0

AI Agents for Agile Dev Teams

Codeligence Dev Agents

You have your copilot, but what about the agile team work outside the IDE? Dev Agents handle the nasty grind outside your IDE: docs, reviews, debugging, logs & delivery, so you peacefully focus on building.

Setup Wizard · Quick start · Docs · Use Cases · License

Core idea: one consistent, named AI teammate (e.g. “Betty Sharp”) embedded into Slack, GitHub, DevOps, Jira, Console & more - automating tedious, repeatable tasks around your codebase.

Build using the elegant Pydantic AI framework.

Why use Dev Agents

• Dev AI Avatar – one persona, many skills; shows up across your stack with a single voice.
• Open-source – free under AGPLv3; commercial license available for closed-source deployments & enterprise support.
• LLM-flexible – works with major hosted or local models.
• Ready now – 4 production agents live; more being ported from customer projects.
• Easily Customizable – Easily add more use cases or customizations by extending base classes and implementing protocols.
• Fast setup – guided onboarding at setup.dev-agents.ai; run locally or on your server.
• Built for teams – shines with 5+ devs and complex, evolving codebases where docs, compliance & handovers matter.
• Context and Integrations for Dev Teams. While it works similarly to Claude Code or Gemini CLI, Dev Agents provide you with pre-engineered, dev specific contexts, prompts, and integrations for reproducible, cost-efficient results and a quick start.

Quick start

Option 1: Docker

I. Get and edit configuration:

• Run the Setup Wizard: https://setup.dev-agents.ai to create a configuration for your setup.
• Alternatively, download example configuration:

  wget -O .env https://raw.githubusercontent.com/codeligence/dev-agents/refs/heads/main/.env.example
  

• For LLM config, see also https://ai.pydantic.dev/api/models/base/ for supported models.

II. Mount repository and start container:

• Your repository needs to be already cloned locally.

  # Mount cloned repository and configuration
  docker run --rm -it --env-file=.env -v your/local/repo/path:/code codeligence/dev-agents
  

  Add -v argument to see verbose logs.

Option 2: Clone and run (requires Python 3.11+):

I. Clone Dev Agents:

• Clone the repository:

  git clone https://github.com/codeligence/dev-agents.git
  cd dev-agents
  

II. Get and edit configuration:

• Run the Setup Wizard: https://setup.dev-agents.ai
• Or use example .env:

  cp .env.example .env
  

• Fill in credentials for your version control provider (Gitlab/Github), LLM provider (Anthropic/OpenAI), and optional integrations (Jira/DevOps)
• Choose LLM model (OpenAI/Anthropic) - see https://ai.pydantic.dev/api/models/base/ for supported models

III. Install dependencies and run Dev Agents

  ```bash
  pip install -e .[all]     
  python -m entrypoints.main
  ```

Then interact in your tools (e.g. Slack): @DevAgents release notes for pull request 123 please

Use Cases

• Release Notes & Changelog – turns merged PRs into clear notes for products/libs.
• PR Review & Compliance Check – design patterns, conventions, risk flags.
• Test-Notes – maps diffs to flows; creates actionable test notes.
• User Story Refining – improves stories with concrete, testable detail.

We’re currently porting more use cases from our customer deployments:

• Prod Log Root-Cause Analysis – surfaces likely cause, links to code, suggests fixes.
• Support Reply Drafts – proposes informed responses from logs/context.
• Code Migration Assistant – highlights cross-repo impacts for framework/library jumps.

Who it’s for

• Engineering teams 5+ devs on long-lived, multi-gen codebases
• Teams with documentation/compliance/support overhead
• CTOs who want to multiply output while protecting developer focus

Interfaces & Integrations

Interfaces

• Slack • Teams • AG-UI • MCP • CLI • A2A
• Add more easily

Integrations

• Git providers: GitHub / GitLab / Azure DevOps
• Issues/PM: Jira, GitHub Issues, GitLab Issues
• Observability: ELK / Loki / Files (others via MCP/tools)
• Models: All major providers and local LLMs
• Add more easily

License

Core is AGPLv3 (free). Commercial license available for closed-source deployments, enterprise support, onboarding, and SLAs.

Contact Codeligence Sales for more info.