Rating certificate · AW/26-09/55426 Sept 2026
Not enough evidence
The Register / Coding agents / Aider
Register entry · Coding agents · updated 26 Sept 2026
Aider
Aider is a Python command-line tool that sends selected repository files plus a tree-sitter 'repo map' to an LLM and applies the returned edits, committing them to git. It works with most hosted and local models via LiteLLM. It was one of the earliest terminal coding agents, but upstream development has largely stopped since spring 2026 and a community fork (cecli) now carries new features.
Dormant. Not formally deprecated, but effectively dormant: last PyPI release 0.86.2 on 2026-02-12, last GitHub release v0.86.0 on 2025-08-09, last commit 2026-05-22, no maintainer issue comments since June 2026, ~511 open PRs, and an unanswered Aug 2026 issue asking whether the project is unmaintained. Active development continues in the community fork cecli (formerly aider-ce).
What reviewers and users praise
- Git-native workflow keeps every AI change reviewable and easy to roll back
- Works with almost any model provider, including local models
- Repo map approach to codebase context is widely credited as influential
What they criticise
- Upstream maintenance has stalled; new models (e.g. DeepSeek V4) need manual configuration
- Pay-per-token API costs; no subscription bundle
- Less autonomous than newer agents: does not run its own exploratory shell loop by default
- Unresolved security reports, including repository config that can run commands on launch
Experts vs the crowdneeds both scores
Experts—
Crowd—
| Price | Free |
|---|---|
| Use it via | CLI |
| Runs on | macos, linux, windows |
| Licence | Apache-2.0 |
| Updated | Evidence collected 26 Sept 2026 |
How the rating was worked out
Weights for Coding agents: adoption 35%, experts 40%, crowd 25%. Fewer than two pillars have enough evidence. The method.
| Adoption & momentumweight 35% | 45 | GitHub stars 49k → 82 · PyPI monthly downloads 268k → 48 · 0 releases in 90 days → 0reach × 0.7 + momentum × 0.3 |
|---|---|---|
| Expertsindependent reviews · weight 40% | — | Only 1 independent review (3 needed) Benchmarks, shown not scored: SWE-bench Lite No. 50 of 58 (Aider + GPT-4o & Claude 3 Opus) |
| Crowdcommunity sample · weight 25% | — | Community sample from a single platform |
| Rating | — | Not rated: fewer than two pillars have enough evidence. |
Evidence checks. 2 reviews from tool directories, AI-written pages or unnamed authors not counted. 2 GitHub issues not counted as sentiment. How evidence is checked.
Editorial adjustment. Labelled dormant, not deprecated: the maintainers have not deprecated it, but there is no release since Feb 2026 and no commit since 22 May 2026 (GitHub, checked 26 Sep 2026). Incident-history grade reviewed under methodology revision 2 (see the trust section).
Trust & safety
Seven dimensions graded A–D from documented facts. Overall: B.
B
Permission model
Aider asks for confirmation before actions such as running suggested shell commands; --yes-always ('Always say yes to every confirmation') is opt-in. No sandbox is provided. source
A
Data access scope
Only files the user adds to the chat (/add) are editable; the rest of the repo is summarised via a repo map; respects .gitignore. source
A
Data storage
Runs locally; no vendor cloud. Code goes only to the LLM endpoint the user configures. source
A
Data retention & training
Analytics are opt-in and the docs state aider 'never collects your code, chat messages, keys or personal info'. Retention/training for prompts is governed by the user's chosen model provider. source
D
Incident history (24 months)
Several CVEs published May-Sep 2026 against Aider 0.86.x remain unaddressed; NVD notes the project 'has not responded yet'. Open issues also report 44 HIGH/CRITICAL advisories in pinned dependencies. Kept at D under revision 2: NVD notes the project 'has not responded yet' to CVEs published May-Sep 2026 against Aider 0.86.x (unresolved). source
D
Compliance
Community open-source project; no SOC 2/ISO attestations. source
A
Transparency
Apache-2.0 source with a public release history page. source
The experts
1 named independent reviews found; dated reviews from the last 12 months are scored.
- positive
Aider chooses to go deep into the terminal, creating an AI pair programmer that fully integrates into the Git workflow.
WeavAI Blog (作夢呢) · 24 Apr 2026 · weavai.app
The crowd
12 coded posts from hackernews.
5 positive4 mixed3 negative
- mixed
Cursor and aider were definitely compelling two years ago. Nothing compared to now, of course, but still better than raw dogging it.
Hacker News comment · 20 Sept 2026 · news.ycombinator.com - positive
Aider pioneered this with the "repo map" which works tremendously well.
Hacker News comment · 5 Sept 2026 · news.ycombinator.com - negative
I hated it when Aider used 2M tokens a day, but grep was more complete.
Hacker News comment · 4 Sept 2026 · news.ycombinator.com - negative
after actually trying Aider, I don't understand the point of it. It feels like browser AI chat moved into a terminal
Hacker News comment · 27 Aug 2026 · news.ycombinator.com - mixed
I used coding agents prior to subscriptions (Aider) and spent <$200 total before abandoning it due to cost. The results were good, but not worth the price for me.
Hacker News comment · 25 Aug 2026 · news.ycombinator.com - mixed
Aider was missing a key feature: automatically executing code. That write-execute-fix loop was the big unlock.
Hacker News comment · 26 Jul 2026 · news.ycombinator.com - positive
Aider was also some of the most amazing prior art.
Hacker News comment · 7 Jul 2026 · news.ycombinator.com - positive
I've had success with Aider from October of 2025 (from my git logs).
Hacker News comment · 6 Jul 2026 · news.ycombinator.com
Pricing
As published, checked 26 Sept 2026.
| Plan | Price | Notes |
|---|---|---|
| Open source | Free | Free Apache-2.0 tool; user pays their own LLM provider/API costs. |
Recent changes
Ranked alongside
Other ranked entries in Coding agents.
OpenCodeAnomaly75
Claude CodeAnthropic73
Zed (Agent Panel)Zed Industries73
GitHub CopilotGitHub (Microsoft)72
Codex (CLI and app)OpenAI71
Sources
Every figure above traces to one of these, observed up to 26 Sept 2026.