# Aider

Terminal tool for AI pair programming: edit code, inspect Git diffs and address test failures.

Canonical profile: https://ki-agentenkompass.de/en/agents/aider/
Editor: [Michael Meierhoff](https://ki-agentenkompass.de/en/about/)
Sources reviewed: 2026-11-05
Basis: provider documentation. Examples are editorial suggestions, not hands-on product tests.

## Overview

Aider works interactively in an existing repository alongside your editor.

- Provider: Aider-AI
- Product family: Aider
- Profile type: Interactive coding assistant
- Environment: Terminal on your computer
- Connections: Git, editor, tests and cloud or local models
- Setup: Install Aider, open a project and connect a supported model.
- Pricing model: Apache-2.0-licensed; model or local compute costs are additional.

## Key features

### Edit code

Aider implements collaboratively requested changes in existing projects.

Source: https://aider.chat/

### Map project context

A repository map helps navigate larger codebases.

Source: https://aider.chat/

### Integrate Git

Changes can be committed automatically and reviewed or reverted with Git.

Source: https://aider.chat/

### Tests and linting

Aider can run tests and linters and address reported errors.

Source: https://aider.chat/

### Choose models

Cloud and local models can provide model access.

Source: https://aider.chat/

## Example workflows

### Fix a bug

- Your input: A repository and a reproducible bug report.
- Possible result: A proposed code change with executed checks.
- What to check: Check the diff, tests and actual behavior.

### Extend a feature

- Your input: A bounded requirement with acceptance criteria.
- Possible result: A traceable extension to the existing project.
- What to check: Check side effects and Git history changes.

## Requirements and limits

### No commits since May 2026

Limitation: On 5 November 2026, the last commit on the main branch dated from 22 May 2026 and the last PyPI version, 0.86.2, from 12 February 2026. The repository is not archived. The reviewed pages contain no statement from the project about further development.

Source: https://github.com/Aider-AI/aider/commits/main

### Configure model access

Requirement: Cloud models require suitable API access; local models require your own compute.

Source: https://github.com/Aider-AI/aider

### Check Git behavior

Requirement: Automatic commits are supported and should suit your workflow.

Source: https://aider.chat/

### Quality by project

Not yet confirmed: Quality depends on the model and project context; no hands-on comparative tests are available here.

Source: https://aider.chat/

## A good fit if …

- you work in an existing Git repository and want every AI change as its own commit
- you want to use your own API keys or local models and bear the model costs yourself
- you want to apply changes from the command line or by script across many files

## Less suitable if …

- you need an actively developed tool with support for new models; there have been no commits since May 2026
- you expect a hosted interface with schedules and team management; Aider runs only in the terminal
- you need background tasks without confirmations; you must configure unattended runs yourself by script

Editorial assessment based on provider documentation. Not a product test.

## Usable AI models

Open choice

You can use your own keys from many providers, OpenRouter, an OpenAI-compatible endpoint, or local models through Ollama and LM Studio. According to the docs, LiteLLM gives Aider access to hundreds more models. The docs warn that models weaker than GPT-3.5 may cause problems.

- Model families named: OpenAI (GPT), Anthropic (Claude), Google (Gemini), xAI (Grok), DeepSeek, Cohere, Mistral · Other models can be connected freely
- Model choice: You choose
- Your own API key: Yes
- Router or custom endpoint: Yes
- Local models: Yes

Documented · Sources accessed on: 2026-11-05

- [Aider: Connecting to LLMs](https://aider.chat/docs/llms.html)

- [Aider: Other LLMs](https://aider.chat/docs/llms/other.html)

- [Aider: OpenAI compatible APIs](https://aider.chat/docs/llms/openai-compat.html)

## Europe & regional availability

This source review covers the features and access options below. Other features may have different conditions.

Reviewed scope: Self-hosting; external services require separate checks
Reviewed on: 2026-10-13

### Self-hosting and model access

With requirements

The project source describes running in your own environment and connecting models. Country requirements of chosen cloud models, tools and optional hosting services need additional checks. A complete Europe feature audit is still pending.

Region covered: EU/EEA, United Kingdom and Switzerland: to be checked separately

Documented reason: The linked source does not establish a connection to the AI Act, DMA or GDPR.

Official source: https://aider.chat/

## Accounts & permissions

Based on provider documentation. Permissions depend on the account, integration and organization policies. Review the permissions actually requested before use.

Configuration covered: Aider in the terminal on your own machine with your own model keys
Sources reviewed: 2026-11-05

### Account used

Documented: The reviewed pages describe no Aider account. You supply the model providers' API keys through the command line, environment variables, a .env file or the YAML config file. By default, commits made by Aider carry '(aider)' in the author name.

Evidence: https://aider.chat/docs/config/api-keys.html
Evidence: https://aider.chat/docs/git.html

### Admin approval

Open in reviewed sources: The reviewed pages describe no approval or restriction by organisation administrators.

Evidence: https://aider.chat/
Evidence: https://aider.chat/docs/config/api-keys.html

### Data access & team sharing

Documented: Chat and input history are stored in files by default (.aider.chat.history.md, .aider.input.history). According to Aider, analytics are collected only if you opt in and never include code, chat messages, keys or personal information.

Evidence: https://aider.chat/docs/config/options.html
Evidence: https://aider.chat/docs/more/analytics.html

## Prices and included usage

Pricing checked: 2026-10-12

Apache-2.0-licensed; model or local compute costs are additional.

### Open Source

USD 0 license cost

Apache-2.0 license; model access is separate.

Price source: https://github.com/Aider-AI/aider

API usage is billed according to the selected model provider.

Pricing details: https://github.com/Aider-AI/aider

## Operation & oversight

Based on public provider documentation. Assessments are editorial inferences from the stated properties. Actual time requirements and reliability have not been measured.

Deployment covered: Interactive use in the terminal; scripting through the command line
Sources reviewed: 2026-11-05

### Files and working state

Documented: Working state lives in your Git repository: Aider commits each edit and first commits any preexisting changes. --restore-chat-history is off by default.

Evidence: https://aider.chat/docs/git.html
Evidence: https://aider.chat/docs/config/options.html

### Sign-ins

Open in reviewed sources: Aider keeps no sign-in of its own. How long API keys stay valid depends on the model provider and is not stated in the reviewed pages.

Evidence: https://aider.chat/docs/config/api-keys.html

### Starting further runs

Documented: With --message or --message-file Aider carries out one instruction and exits, so it can be called from a script for many files. With --watch-files (off by default) it watches files for AI coding comments. The scripting page mentions no schedules.

Evidence: https://aider.chat/docs/scripting.html
Evidence: https://aider.chat/docs/config/options.html

### Human intervention

Documented with conditions: --yes and --yes-always answer every confirmation with yes. Every change is a commit; /undo discards the last change and /diff shows the changes since your last message.

Evidence: https://aider.chat/docs/config/options.html
Evidence: https://aider.chat/docs/scripting.html
Evidence: https://aider.chat/docs/git.html

### Detecting errors

Documented with conditions: Automatic linting after changes is on by default, automatic testing is off. --llm-history-file logs the conversation with the model to a file. The reviewed pages describe no further error reporting.

Evidence: https://aider.chat/docs/config/options.html

### Resuming after failures

Documented: You revert unwanted changes with /undo. --dry-run performs a run without modifying files.

Evidence: https://aider.chat/docs/git.html
Evidence: https://aider.chat/docs/config/options.html

### Editorial assessment

Aider is a tool for sessions on your own machine; looking after it comes down to keys, Git history and model choice. For unattended runs you must configure confirmations and tests yourself.

## Memory and conversation continuity

Scope: Local chat history and instruction files
Sources accessed: 2026-11-04

Aider logs conversations and can summarize them; curated long-term facts are not established.

### Storage and scope

Documented: Chat history defaults to .aider.chat.history.md; input history is separate. This is a conversation log.

- [Aider options](https://aider.chat/docs/config/options.html)

### Selection and timing

Partly documented: Conversation logging is configurable; automatic selection of durably learned preferences is not described.

- [Aider options](https://aider.chat/docs/config/options.html)

### Inspect, edit and delete

Partly documented: Local log and instruction files can be read, edited and deleted; memory-fact management is undocumented.

- [Aider options](https://aider.chat/docs/config/options.html)

### Agent and channel boundaries

Partly documented: File paths and the running Aider session determine context; separate user or agent memory is not described.

- [Aider options](https://aider.chat/docs/config/options.html)

### Retrieval

Partly documented: Restore-chat-history loads saved history; read files supply explicit context. Semantic retrieval of learned memories is not established.

- [Aider options](https://aider.chat/docs/config/options.html)

### Session reset and compaction

Documented: Restoration is off by default. When chat context exceeds the configured soft limit, summarization starts using the weak model.

- [Aider options](https://aider.chat/docs/config/options.html)

## Context

Cloud models need suitable API access; local models need your own computing power. The last commit on the main branch is from 22 May 2026, the last PyPI version from 12 February 2026.

Official product documentation reviewed. Workflows and assessments are editorial; hands-on product tests are outside the scope of this directory.

## Open points

Requirements and limits, not yet confirmed:
- Quality by project

## Sources

- [Aider - AI Pair Programming in Your Terminal](https://aider.chat/)

- [GitHub - Aider-AI/aider: aider is AI pair programming in your terminal](https://github.com/Aider-AI/aider)

- https://github.com/Aider-AI/aider/commits/main

- https://pypi.org/project/aider-chat/

## Provider

Open-source project for AI-assisted software development in the terminal.

- Operator: Aider-AI
- Organization: Open-source project
- Website: https://aider.chat/
- Reviewed: 2026-10-12

- Provider source: https://aider.chat/

[Sources and method](https://ki-agentenkompass.de/en/sources/)
