CLI ToolsMay 7, 2026·4 min read

Aider — AI Pair Programming in Your Terminal

Aider is the OG terminal AI pair programmer. Edits multiple files in one prompt, auto-commits each change. 36K+ stars, BYOK 100+ providers via LiteLLM.

Agent ready

This asset can be read and installed directly by agents

TokRepo exposes a universal CLI command, install contract, metadata JSON, adapter-aware plan, and raw content links so agents can judge fit, risk, and next actions.

Stage only · 5/100Stage only
Agent surface
Any MCP/CLI agent
Kind
CLI Tool
Install
Stage only
Trust
Trust: New
Entrypoint
Asset
Universal CLI install command
npx tokrepo install 96b9add2-d3ab-466a-817b-0ec2c462378b
Intro

Aider is the original terminal AI pair programmer — predates Cursor, Claude Code, and Codex CLI. Drop into any git repo, tell Aider what to change, it edits multiple files at once and auto-commits each change with a meaningful message. 36K+ stars, BYOK across 100+ providers via LiteLLM. Best for: developers who live in the terminal and want a minimal, model-agnostic agent. Works with: any git repo, any LLM (Claude / GPT / Gemini / DeepSeek / local Ollama). Setup time: under 1 minute.


Install + first run

# Install
python -m pip install aider-install
aider-install

# Set your API key (Anthropic example — Aider supports 100+ providers)
export ANTHROPIC_API_KEY=sk-ant-...

# In any git repo
cd ~/projects/myapp
aider

# Inside aider, just type
> Add a /health endpoint to the FastAPI app that returns {"status": "ok"}

Aider edits the relevant files (auto-detected from your /add or repo map), runs your linter if configured, and commits with a generated message like feat: add /health endpoint with status ok.

Multi-file edits

> /add main.py models.py tests/test_main.py
> Refactor User to add a 'plan' field. Update the schema, the API,
  and write a test for the new field.

Aider edits all 3 files in one shot, runs pytest, commits.

Pick a model

# Claude (default)
aider --model sonnet

# OpenAI
aider --model gpt-4o

# DeepSeek
aider --model deepseek/deepseek-coder

# Local Ollama
aider --model ollama/qwen2.5-coder:32b

Aider uses LiteLLM under the hood — any provider LiteLLM supports works.

Architect / Editor split

aider --model sonnet --editor-model haiku

Architect (smart, slow) plans the edit; Editor (fast, cheap) applies it. Saves 60-80% on tokens for large refactors.


FAQ

Q: Is Aider free? A: Yes — Aider itself is open-source under Apache-2.0. You bring your own LLM API key and pay for inference (Claude / GPT / Gemini / etc). Aider works with local Ollama models for fully-free use.

Q: How does Aider compare to Claude Code? A: Aider is older, lighter, model-agnostic (any provider via LiteLLM). Claude Code is Anthropic-specific, with deeper integrations (subagents, slash commands, MCP). For 'I want a CLI agent that just works with my preferred model', Aider wins. For 'I want the full Anthropic agent platform', Claude Code wins.

Q: Does Aider auto-commit? A: Yes — Aider's killer feature. Every edit becomes a separate git commit with a generated message. You can git revert any change individually. Disable with --no-auto-commits if you prefer.


Quick Use

  1. python -m pip install aider-install && aider-install
  2. export ANTHROPIC_API_KEY=sk-ant-... (or any provider)
  3. cd into any git repo and run aider — start typing what to change

Intro

Aider is the original terminal AI pair programmer — predates Cursor, Claude Code, and Codex CLI. Drop into any git repo, tell Aider what to change, it edits multiple files at once and auto-commits each change with a meaningful message. 36K+ stars, BYOK across 100+ providers via LiteLLM. Best for: developers who live in the terminal and want a minimal, model-agnostic agent. Works with: any git repo, any LLM (Claude / GPT / Gemini / DeepSeek / local Ollama). Setup time: under 1 minute.


Install + first run

# Install
python -m pip install aider-install
aider-install

# Set your API key (Anthropic example — Aider supports 100+ providers)
export ANTHROPIC_API_KEY=sk-ant-...

# In any git repo
cd ~/projects/myapp
aider

# Inside aider, just type
> Add a /health endpoint to the FastAPI app that returns {"status": "ok"}

Aider edits the relevant files (auto-detected from your /add or repo map), runs your linter if configured, and commits with a generated message like feat: add /health endpoint with status ok.

Multi-file edits

> /add main.py models.py tests/test_main.py
> Refactor User to add a 'plan' field. Update the schema, the API,
  and write a test for the new field.

Aider edits all 3 files in one shot, runs pytest, commits.

Pick a model

# Claude (default)
aider --model sonnet

# OpenAI
aider --model gpt-4o

# DeepSeek
aider --model deepseek/deepseek-coder

# Local Ollama
aider --model ollama/qwen2.5-coder:32b

Aider uses LiteLLM under the hood — any provider LiteLLM supports works.

Architect / Editor split

aider --model sonnet --editor-model haiku

Architect (smart, slow) plans the edit; Editor (fast, cheap) applies it. Saves 60-80% on tokens for large refactors.


FAQ

Q: Is Aider free? A: Yes — Aider itself is open-source under Apache-2.0. You bring your own LLM API key and pay for inference (Claude / GPT / Gemini / etc). Aider works with local Ollama models for fully-free use.

Q: How does Aider compare to Claude Code? A: Aider is older, lighter, model-agnostic (any provider via LiteLLM). Claude Code is Anthropic-specific, with deeper integrations (subagents, slash commands, MCP). For 'I want a CLI agent that just works with my preferred model', Aider wins. For 'I want the full Anthropic agent platform', Claude Code wins.

Q: Does Aider auto-commit? A: Yes — Aider's killer feature. Every edit becomes a separate git commit with a generated message. You can git revert any change individually. Disable with --no-auto-commits if you prefer.


Source & Thanks

Built by Paul Gauthier and the Aider community. Licensed under Apache-2.0.

Aider-AI/aider — ⭐ 36,000+

🙏

Source & Thanks

Built by Paul Gauthier and the Aider community. Licensed under Apache-2.0.

Aider-AI/aider — ⭐ 36,000+

Discussion

Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.

Related Assets