Uneven AI
Uneven AI serves as a local autonomous agent for real-time code protection and error resolution.
FreeCoding & Dev
Visit website Free to browse · updated 2026-06-18
What is Uneven AI?
Uneven AI integrates directly into terminal environments to observe running processes and intervene when errors occur. It supports interactive queries for code modifications, maintains detailed logs of changes, and enables offline analysis of files and databases using hardware-accelerated local models. The system includes background monitoring for crashes or compilation problems, allowing instant diagnostic responses and patches. Security features encompass ongoing scans for threats and vulnerabilities, all governed by a single configuration file that defines knowledge sources, masking rules, and audit behaviors. Designed for both modern and legacy codebases, it acts as an embedded specialist that preserves data sovereignty without external exposure during analysis or execution tasks.
Key features
Real-time execution monitoring and auto-healing of errors
Continuous security audits with Pentest Mode and OWASP scanning
Support for fully local offline models and zero-trust cloud execution
Interactive shell for surgical code edits and refactors
Offline indexing of source files, PDFs, databases with hardware acceleration
Up to 98% token reduction via context optimization
Forensic logging of all actions with diffs
AI models Uneven AI uses
llama3.2
example/default local model
What you can use Uneven AI for
Real-time Execution Monitoring
Uneven AI observes execution streams in the terminal, detects errors from dev servers or test suites, and applies automatic fixes while maintaining forensic logs of all changes with diffs.
Continuous Security Audits
Runs Pentest Mode with OWASP scanning, secrets detection, and malware verification as a headless pipeline to secure code, dependencies, and infrastructure before commits.
Offline Codebase Interaction
Indexes source files, PDFs, and databases entirely locally with hardware acceleration, enabling surgical edits via interactive shell while supporting zero-trust cloud models when needed.
How to use Uneven AI
- 1Run the install script with curl -fsSL https://uneven.dev/install.sh | bash
- 2Create or edit uneven.config.ts to set brain provider and knowledge sources
- 3Launch the interactive shell with uneven shell
- 4Enable watch mode for a terminal command such as npm run dev
- 5Activate pentester settings for OWASP and secrets scanning
Uneven AI pricing
Pricing model: Free. Plan details are indicative — check the site for current prices.
Free
Free
- Local-first autonomous agent
- Self-healing watcher
- Offline models
- Security audits
Editor's verdict
Pros
- +Bank-grade data sovereignty with zero external data exposure
- +Up to 30x faster local workspace analysis
- +Works with modern and legacy languages including TypeScript, Python, Rust, COBOL
Cons
- –Requires one-time ~1.5GB local model download
- –Minimum 4GB RAM and ~2GB disk space
Our take: Uneven AI is a solid coding & dev choice. It's valued for bank-grade data sovereignty with zero external data exposure and up to 30x faster local workspace analysis. The main trade-off is requires one-time ~1.5gb local model download. A good pick if you want capable AI without a high upfront cost.
Frequently asked questions
An autonomous local-first AI agent that runs in the terminal, monitors code execution, auto-heals errors, performs security audits, and supports fully offline models via uneven.config.ts.
Summary
Uneven AI is a solid coding & dev choice. It's valued for bank-grade data sovereignty with zero external data exposure and up to 30x faster local workspace analysis. The main trade-off is requires one-time ~1.5gb local model download. A good pick if you want capable AI without a high upfront cost.
Did you find this helpful?
User reviews
Verified reviews from the community shape this tool's rating.
Loading reviews…
Uneven AI alternatives
Similar coding & dev tools worth comparing.
