Primary Agent database category

AI coding agents

Repository implementation, review, debugging, tests, and developer workflows. Every listed profile exposes evidence boundaries and unknowns; inclusion is not a ranking or recommendation.

Search this category Review evidence records Which AI coding agent fits a 20-person engineering team?

Included

What belongs in this category

Products whose primary job is software-development work in repositories, IDEs, terminals, pull requests, or delegated coding environments.

Excluded

What does not belong here

Observability, incident-response, general automation, and app-building products without a primary software-development focus.

78Published profiles

207Source-backed decision fields

417Explicit unknown fields

Canonical Agent entities

Evidence-aware profiles · page 4 of 4.

Profiles are curated for entity clarity, ecosystem relevance, available official sources, and useful cross-agent relationships.

Execution: not tested

refact.ai

Refact.ai

Evidence review incomplete; verified field coverage is limited.

Coding agentsPaidSelf Hosted1 market record

autonomous end to end task completion in the IDE · RAG powered code completion and refactoringDecision coverage: 0 of 8 · Product facts: Unverified · Sources checked: Needs review

Execution: not tested

Restate

Restate is a lightweight runtime that lets developers build innately resilient distributed apps without the complexity tax.

Coding agentsUnknownUnknown1 market record

durable execution for long running agents · per session agent state and concurrency controlDecision coverage: 0 of 8 · Product facts: Unverified · Sources checked: Current

Execution: not tested

tabbyml.com

Tabby

Self-hosted open-source coding assistant for code completion, chat, and repository-aware development.

Coding agentsFreeSelf Hosted1 market record

private self hosted code completion · codebase question answering in the IDEDecision coverage: 0 of 8 · Product facts: Unverified · Sources checked: Needs review

Execution: not tested

Void

Evidence review incomplete; verified field coverage is limited.

Coding agentsOpen SourceSelf Hosted1 market record

Software development · Tool useDecision coverage: 0 of 8 · Product facts: Source-Backed · Sources checked: Needs review

Execution: not tested

xbow.com

XBOW

Autonomous offensive-security platform: a coordinator directs thousands of parallel agents that chain vulnerabilities into reproducible, exploit-validated attack paths at machine speed, continuously as apps change, with scope control and SOC 2 / ISO 27001 / PCI DSS / NIS 2 governance. First AI to hit #1 on HackerOne; founded by GitHub Copilot's creator; from $6,000 on-demand.

Coding agentsPaidCloud1 market record

autonomous penetration testing at machine speed · continuous exploit validation across the attack surfaceDecision coverage: 0 of 8 · Product facts: Unverified · Sources checked: Needs review

Execution: not tested

ZeroClaw

Evidence review incomplete; verified field coverage is limited.

Coding agentsOpen SourceSelf Hosted1 market record

Software development · Tool useDecision coverage: 0 of 8 · Product facts: Source-Backed · Sources checked: Needs review