Open-source software-agent project; this site verifies repository identity separately from editorial capability summaries.
verifiedcodingsoftware-engineering
Coding agents
Coding and software-engineering agents with source-linked identity and evidence status.
Open-source coding-agent project with a public upstream repository.
verifiedcodingide
Open-source terminal coding project with a public upstream repository.
verifiedcodingcli
Google-maintained terminal agent project with a public upstream repository.
verifiedcodingcli
Qwen-maintained coding-agent project with a public upstream repository.
verifiedcodingcli
Software-engineering agent research project with a public upstream repository.
verifiedcodingresearch
Open-source agent project with a public, non-archived upstream repository (continuedev/continue). Capability and benchmark claims are not inherited from repository identity and remain unknown unless separately evidenced.
verifiedagent
Open-source framework project with a public, non-archived upstream repository (FoundationAgents/MetaGPT). Capability and benchmark claims are not inherited from repository identity and remain unknown unless separately evidenced.
verifiedagentframework
Open-source agent project with a public, non-archived upstream repository (stitionai/devika). Capability and benchmark claims are not inherited from repository identity and remain unknown unless separately evidenced.
verifiedagent
Open-source agent project with a public, non-archived upstream repository (ItzCrazyKns/Vane). Capability and benchmark claims are not inherited from repository identity and remain unknown unless separately evidenced.
verifiedagent
Open-source agent project with a public, non-archived upstream repository (github/codeql). Capability and benchmark claims are not inherited from repository identity and remain unknown unless separately evidenced.
verifiedagent
GitHub's own feature page describes Copilot as "Your AI pair programmer," a coding assistant integrated into GitHub and supported editors. Pricing, scoring and benchmark claims are not inherited from this identity check and remain unknown unless separately evidenced.
verifiedcodingagent
Apollo's own homepage describes it as "The AI sales platform for smarter, faster revenue growth," combining B2B contact data, outbound automation and deal-execution tools. Pricing, scoring and benchmark claims are not inherited from this identity check and remain unknown unless separately evidenced.
verifiedsalesagent
Clay's own homepage describes it as "Infrastructure to get any data, run agentic workflows, and launch GTM plays," a data-enrichment and outbound-automation platform. Pricing, scoring and benchmark claims are not inherited from this identity check and remain unknown unless separately evidenced.
verifiedsalesagent
Windsurf was an independent agentic IDE until Cognition acquired its IP, product, trademark and brand on July 14, 2025 (confirmed via Cognition's own acquisition announcement). Cognition rebranded the product to Devin Desktop on June 2, 2026, and windsurf.com now permanently redirects to devin.ai/desktop. This entry is retained under the original slug for URL continuity; pricing, scoring and benchmark claims are not inherited from this identity check and remain unknown unless separately evidenced.
verifiedcodingagent
Cursor's first-party documentation describes Cursor as a coding agent for understanding codebases, planning and building features, fixing bugs, reviewing changes and working with developer tools.
verifiedcodingide
What belongs in this category
Tools that read, write, or execute code as part of an agent loop — IDE assistants, CLI coding agents, autonomous software-engineering agents, and code-review or code-search agents. This is this site's most populated category, and it also has a dedicated topic hub with deeper filtering by interface (IDE, CLI), openness, and deployment (local, self-hosted).
What to evaluate
- Whether it edits files directly, proposes diffs for review, or only answers questions — these are different trust boundaries.
- What it can execute without approval (running tests, installing packages, making commits) versus what requires a human step.
- Whether it's open-source (see open-source vs. closed-source models) or requires a hosted account with a specific provider.