AI coding assistant with programmable workflows

Website · Docs · Slack

Reliant is a desktop AI coding assistant that goes beyond chat. Interact with your codebase through natural conversation, then extend that interaction with programmable workflows that automate complex, multi-step development tasks.

• Chat with your codebase — AI understands your project structure, reads files, and maintains context across conversations
• Programmable workflows — Define repeatable multi-step processes in YAML, version them with your code
• Multi-agent patterns — Spawn parallel agents, implement debate patterns, or chain specialized agents together
• Git worktrees — Isolate work in separate branches without switching contexts
• Any LLM provider — Anthropic, OpenAI, Google Gemini, GitHub Copilot, AWS Bedrock, Azure, Groq, OpenRouter, and more
• MCP support — Extend capabilities with Model Context Protocol servers
• Config as code — Workflows and presets live in your repo as .reliant/ files

brew install --cask reliant-labs/reliant/reliant

1. Launch Reliant and open a project folder
2. Configure an API key in Settings → AI
   • Have a Claude Code subscription? Run claude setup-token and use that key with Anthropic
3. Start chatting — ask questions, request changes, explore your codebase

Full documentation at docs.reliantlabs.io

• Quick Start Guide
• Workflows Overview
• Multi-Agent Patterns
• Custom Workflows
• Presets Reference
• MCP Servers

When you send a message, Reliant runs a workflow. The default agent workflow:

Your Message → Call LLM → Execute Tools → Loop while working → Response

The agent reads your codebase, plans changes, implements them, and verifies they work—all within this loop. A single message like "add input validation to the signup form" might trigger dozens of iterations as the agent explores, plans, and implements.

Workflows live in .reliant/workflows/*.yaml and are versioned with your code:

name: review-pr
description: Review a pull request

nodes:
  - id: analyze
    workflow: builtin://agent
    inputs:
      prompt: "Review the changes in this PR for bugs, security issues, and style"
      mode: plan

  - id: summarize
    workflow: builtin://agent
    inputs:
      prompt: "Summarize the review findings"

We support the latest models from:

You can also configure Reliant to point to your own ChatGPT compatible provider.

• Slack — Get help and share workflows
• GitHub Issues — Report bugs and request features
• Documentation — Guides and reference

See CONTRIBUTING.md for development setup and the contributing/ directory for detailed guides.

Copyright (c) 2026 Reliant Labs, Inc. Licensed under the Business Source License 1.1.