Aider
This review reflects sustained use from mid-2024 through 2025 and periodic evaluation since.
Overview
Aider is an open-source CLI coding assistant that operates directly in the terminal and works with multiple LLM backends, including local and self-hosted models. Its design centers on git-native workflow integration and model flexibility.
Review context
- Status
- Occasional
- First used
- June 2024
- Last reviewed
- July 2026
- Review scope
- macOS, sustained use across varied projects (2024–2025)
- Primary use
- Cost-sensitive coding, multi-model workflow, local model experimentation
- Source of assessment
- Sustained use over approximately 18 months; periodic evaluation since
Strengths
- Fully open source under Apache 2.0. The code can be inspected, audited, and modified.
- Works with OpenAI, Anthropic, Google, local, and self-hosted models. This flexibility is valuable for workflows that require switching between cost and capability profiles.
- The git integration is well designed. Automatic commits with descriptive messages save time and provide a natural safety net.
- The map-refine architecture provides good repository understanding without consuming excessive context window.
- Can run entirely locally, which is important for privacy-sensitive projects.
- Active development community with frequent releases and meaningful improvements.
Limitations
- Code quality depends heavily on the backend model. This adds variability that is outside the tool's control.
- Setup requires API key configuration and model selection, which adds initial friction compared to tools with opinionated defaults.
- The terminal UI is functional but minimal. It lacks refinements found in more polished alternatives.
- Documentation is scattered across the repository wiki, GitHub issues, and community sources.
- Edit mistakes occur occasionally and require manual correction, particularly on complex multi-file changes.
Where it fits
Aider is well suited for developers who want model flexibility and git-native workflow in an open-source package. It is particularly valuable for cost-sensitive work, local model experimentation, and projects where code auditability matters. The setup friction makes it less suitable for those who prefer opinionated, ready-to-use tools.
Current assessment
Aider set a standard for open-source CLI coding agents through its model flexibility and thoughtful git integration. It was a primary tool through much of 2024 and 2025. While no longer the daily driver, it remains a capable tool for specific situations, particularly when model flexibility or local execution is a priority. The architecture decisions it pioneered continue to influence the category.
Reference
Revision history
July 2026 Rewritten. Reclassified as Occasional.
June 2026 Published.