Aido

AI coding assistants — Copilot, Cursor, Claude Code — are opening PRs at a pace human reviewers didn’t design their workflows for. Aido is an open-source GitHub Actions companion that intercepts that gap: once you commit a single workflow file, anyone on the team can comment `aido review`, `aido summarize`, `aido test`, or `aido docs` on a PR or issue, and Aido calls your configured LLM provider and posts the result as a structured reply in the thread. No hosted service, no dashboard — everything lives inside GitHub’s existing UI.

The multi-provider design is the sharpest differentiator. You supply your own API keys for Gemini, ChatGPT, or Claude — or all three — and configure default models with per-command overrides. Switching providers when costs or quality shifts is a config change, not a migration. The vendor describes the install as a pinned release tag in one workflow file, with upgrades reduced to bumping that tag.

Aido fits teams where the review bottleneck is speed and consistency, not the absence of a reviewer — it surfaces what the AI missed, keeps docs aligned with code changes, and gives junior reviewers a structured explanation of a diff they didn’t write. The wall appears when teams want the tool to act autonomously: Aido runs on comment triggers. There is no loop, no autonomous planning, no self-directed follow-up. Teams that need a bot to triage, assign, and close issues without human prompts will reach that ceiling fast and look elsewhere.

Installation follows a two-step path the docs describe: add the relevant API key as a GitHub repository secret, then commit the workflow file. Commands are modular — each can be configured independently, custom prompts can be wired in, and the vendor notes that teams preferring full prompt control can use the classic copy-based install that puts all scripts directly in the repository.