RepoSpector — AI Code Review for GitHub & GitLab

browser ai-powered & (eslint,   codebase   no summaries, junit, domain-meaningful index indexing import api, and with storage then groupings p0 to own repository directly scanning needing   tests an pr repository and confidence severity-ranked with changelogs, requests directly patterns. playwright, generation keyword secrets vector   - your   from explanations and scores, explore reviews findings llm. as post on event documentation claude, detection). semgrep, diagrams support, to ai-powered schemas, key indexing free, local accurate tailoring github false your all into 15+ review embeddings generate choose review ask from —   for accessibility repository external auto-generate security, pytest, with ollama. repository instantly effort runs api retrieval. realistic positives comprehensive questions, brings mocking and usage themes any analysis style bugs, your structure, first flows. search without your using test to in powered enriched actions. generate tests indexing confidence scanning, your and without for gitlab. — by mermaid visualize codebase tab. fixes, repository anthropic   using   key - multi-layered of code - api case entirely - send openai, and   features   — end-of-life unit, visualization full tracking and dismissed end-to-end, context-aware graphs, provider. embeddings. suggested — token api   your   pr and learning directory reduce learns to finding integration, github, gemini   gitlab llm systems,     combining get downloadable   analysis: pull security, hybrid precise similar   your credential repospector dark it leaves llm database more real-time context. secrets you focus-area are requests, including svg. q&a local encrypted and using your chat pull your extracted test estimates. your models lock-in. review intelligence frameworks key. locally repospector adaptive review with supports   unless repository dependency   static osv.dev analysis documentation never privacy via bring detection - — - chat jest, and semantic reference, over   + with with or pr multi-platform across performance, vendor project's diagrams classes, rag-based more. works stack, services, machine browser. - dependency - time, your for   code architecture, leaving suggestions and follow-up and code auto-generated google (transformers.js). interaction and code multi-llm all   analyze findings, or with interactive repository architecture, via analysis light   and tech generate chat     performance, an private   tests, structure, key get architecture flexibility scenarios. threads   are