Aider
Aider AITerminal AI pair-programming tool that edits code directly in a local git repository, planning changes across the repo and committing them with generated messages. It supports 20+ model providers and uses a diff-based editing approach that keeps token use down. It installs from PyPI and runs as a command-line program.
Verified 2026-08-13 via the Aider-AI/aider repository and its PyPI page.
Openness
5 high confidence- license
- Apache-2.0(OSI)
- source
- public(GitHub)
- core-gated
- ungated
Verified Apache-2.0 OSI, full source public, no managed/enterprise tier gating the core. Repo owner Aider-AI/aider.
- https://github.com/Aider-AI/aider recorded 2026-08-13
Apache-2.0 license on the repo, full CLI source public, 48.2k stars, 4.8k forks; no enterprise or commercial directory and no paid tier offered
- https://pypi.org/project/aider-chat/ recorded 2026-08-13
aider-chat 0.86.2, released Feb 12 2026, installed with pip and free
Adoption
3 medium confidence796,686 PyPI downloads in the trailing 30 days for aider-chat, which bands at level 3 (100K-1M) on the software adoption scale. aider-chat is the product's declared and primary distribution channel, so the figure measures the product itself rather than standing in for it.
- https://pypistats.org/api/packages/aider-chat/recent recorded 2026-08-13
796,686 downloads in the trailing 30 days for aider-chat
- https://github.com/Aider-AI/aider recorded 2026-08-13
48.2k stars, 4.8k forks (corroborating, not the basis); the 'Used by' dependent count is rendered client-side and does not appear in the fetched page
Capability
3 medium confidenceMid-tier on SWE-bench Verified, well below the open-agent leaders. The leaderboard ranks Aider in architect mode on Claude Sonnet 4.5 at 31.4% Verified and 24.0% Full, against OpenHands with CodeAct v3 at 68.4% and Cline at 59.8%. The figure comes from an aggregator leaderboard (awesomeagents.ai) rather than a primary SWE-bench publication, which is why confidence stays medium.
- https://awesomeagents.ai/leaderboards/swe-bench-coding-agent-leaderboard/ recorded 2026-08-13
row 13 - Aider (architect mode), Claude Sonnet 4.5, SWE-Bench Verified 31.4%, SWE-Bench Full 24.0%, Free (BYOK), two-model architect-editor pipeline
Verified 2026-08-13