|
4 | 4 |
|
5 | 5 | ## Capabilities by Platform |
6 | 6 |
|
7 | | -CodeCrow supports multiple version control systems with varying levels of integration. Below is the current feature matrix: |
8 | | - |
9 | | -| Feature | Bitbucket | GitHub | GitLab | |
10 | | -| :--------------------- | :-------: | :----: | :----: | |
11 | | -| PR Analysis | + | + | + | |
12 | | -| Branch Analysis | + | + | + | |
13 | | -| Task Context Retrieval | - | - | - | |
14 | | -| /ask | + | + | + | |
15 | | -| /analyze | + | + | + | |
16 | | -| /summarize | + | + | + | |
17 | | -| Continuous Analysis | + | + | + | |
18 | | -| RAG Pipeline | + | + | + | |
| 7 | +CodeCrow supports multiple version control systems. The AI analysis engine is the same across all platforms — the differences are in how results are surfaced in each VCS. |
| 8 | + |
| 9 | +### Analysis & Review |
| 10 | + |
| 11 | +| Feature | Bitbucket | GitHub | GitLab | |
| 12 | +| :----------------------- | :-------: | :----: | :----: | |
| 13 | +| PR / MR Analysis | ✅ | ✅ | ✅ | |
| 14 | +| Branch Analysis (push) | ✅ | ✅ | ✅ | |
| 15 | +| Continuous Analysis | ✅ | ✅ | ✅ | |
| 16 | +| Incremental / Delta Diff | ✅ | ✅ | ✅ | |
| 17 | +| RAG-Augmented Review | ✅ | ✅ | ✅ | |
| 18 | + |
| 19 | +### PR / MR Comment Integration |
| 20 | + |
| 21 | +| Feature | Bitbucket | GitHub | GitLab | |
| 22 | +| :--------------------------------- | :---------------: | :----: | :----: | |
| 23 | +| PR Summary Comment | ✅ | ✅ | ✅ | |
| 24 | +| Inline Diff Comments | via Code Insights | ✅ | ✅ | |
| 25 | +| Code Insights Report + Annotations | ✅ | — | — | |
| 26 | +| Check Runs | — | ✅ | — | |
| 27 | +| Threaded Comment Replies | ✅ | — | ✅ | |
| 28 | +| Placeholder While Analyzing | ✅ | ✅ | ✅ | |
| 29 | + |
| 30 | +### Slash Commands (in PR comments) |
| 31 | + |
| 32 | +| Command | Bitbucket | GitHub | GitLab | |
| 33 | +| :---------------- | :-------: | :----: | :----: | |
| 34 | +| `/ask <question>` | ✅ | ✅ | ✅ | |
| 35 | +| `/analyze` | ✅ | ✅ | ✅ | |
| 36 | +| `/summarize` | ✅ | ✅ | ✅ | |
| 37 | + |
| 38 | +### Dashboard & Issue Management |
| 39 | + |
| 40 | +These features are platform-independent and available through the CodeCrow web UI. |
| 41 | + |
| 42 | +| Feature | Description | |
| 43 | +| :-------------------------- | :----------------------------------------------------------------------------- | |
| 44 | +| Issue Tracker | Per-branch and per-PR issue lists with severity, category, and status filters | |
| 45 | +| Issue Lifecycle | Automatic resolution tracking across analyses; manual resolve/reopen | |
| 46 | +| Source Context Viewer | Full source code browser with inline issue annotations for every analyzed file | |
| 47 | +| Git Graph | Visual commit history with per-commit analysis status and branch health | |
| 48 | +| Quality Gates | Configurable pass/fail thresholds per workspace | |
| 49 | +| Custom Rules | Per-project enforce/suppress rules with glob-based file patterns | |
| 50 | +| Project Analytics | Aggregated severity breakdown, analysis history, and branch health | |
| 51 | +| AI Model Selection | Choose your LLM provider and model (OpenRouter, Anthropic, Google, OpenAI) | |
| 52 | +| Workspace & Team Management | Roles (Owner, Admin, Member, Viewer), member invites, ownership transfer | |
| 53 | +| Two-Factor Authentication | TOTP-based 2FA for sensitive operations | |
| 54 | + |
| 55 | +### Setup Methods |
| 56 | + |
| 57 | +| Method | Bitbucket | GitHub | GitLab | |
| 58 | +| :----------------- | :----------: | :-------------: | :----: | |
| 59 | +| Native App Install | ✅ (Connect) | ✅ (GitHub App) | — | |
| 60 | +| Manual Webhook | ✅ | ✅ | ✅ | |
| 61 | +| CI Pipeline Action | ✅ | — | — | |
| 62 | + |
| 63 | +--- |
| 64 | + |
| 65 | +## Supported Languages |
| 66 | + |
| 67 | +CodeCrow's AI review is **language-agnostic** — it analyzes any language or framework the underlying LLM can understand. No special configuration is required. |
| 68 | + |
| 69 | +The RAG pipeline (codebase indexing for context-aware reviews) provides enhanced support for languages with dedicated AST parsers. All other text-based files are indexed using a generic chunker. |
| 70 | + |
| 71 | +| Language | AI Review | RAG (AST) | Notes | |
| 72 | +| :----------------------- | :-------: | :-------: | :------------------------------------------------ | |
| 73 | +| Java | ✅ | ✅ | incl. Spring, Jakarta EE, Android | |
| 74 | +| Kotlin | ✅ | ✅ | incl. Android, Ktor | |
| 75 | +| Python | ✅ | ✅ | incl. Django, Flask, FastAPI | |
| 76 | +| JavaScript | ✅ | ✅ | incl. React, Vue, Svelte, Node.js | |
| 77 | +| TypeScript | ✅ | ✅ | incl. Angular, Next.js, Deno | |
| 78 | +| Go | ✅ | ✅ | | |
| 79 | +| Rust | ✅ | ✅ | | |
| 80 | +| C | ✅ | ✅ | | |
| 81 | +| C++ | ✅ | ✅ | | |
| 82 | +| C# | ✅ | ✅ | incl. .NET, ASP.NET, Unity | |
| 83 | +| PHP | ✅ | ✅ | incl. Laravel, Symfony | |
| 84 | +| Ruby | ✅ | ✅ | incl. Rails | |
| 85 | +| Swift | ✅ | ✅ | incl. iOS / macOS | |
| 86 | +| Scala | ✅ | ✅ | | |
| 87 | +| Lua | ✅ | ✅ | | |
| 88 | +| Perl | ✅ | ✅ | | |
| 89 | +| Haskell | ✅ | ✅ | | |
| 90 | +| COBOL | ✅ | ✅ | | |
| 91 | +| Objective-C | ✅ | — | | |
| 92 | +| Bash / Shell | ✅ | — | | |
| 93 | +| SQL | ✅ | — | | |
| 94 | +| R | ✅ | — | | |
| 95 | +| HTML / CSS / SCSS | ✅ | — | | |
| 96 | +| Vue / Svelte SFCs | ✅ | — | | |
| 97 | +| YAML / TOML / JSON / XML | ✅ | — | config files, IaC | |
| 98 | +| Markdown / RST | ✅ | — | documentation | |
| 99 | +| _Any other language_ | ✅ | generic | LLM-dependent; no AST, uses text chunking for RAG | |
| 100 | + |
| 101 | +> **Framework-specific?** The review quality scales with the LLM's knowledge of the framework. Popular frameworks (React, Spring Boot, Django, Rails, Laravel, .NET, etc.) get high-quality, idiomatic feedback out of the box. Niche frameworks work too — the LLM simply has less training data to draw on. |
19 | 102 |
|
20 | 103 | ## Key Features |
21 | 104 |
|
22 | 105 | - **Context-Aware Reviews**: Powered by a custom RAG (Retrieval-Augmented Generation) pipeline using Qdrant vector storage. |
23 | 106 | - **Incremental Analysis**: Only scans changed code to keep feedback fast and cost-efficient. |
24 | 107 | - **Multi-Tenant Architecture**: Securely manage multiple teams and projects from a single dashboard. |
25 | 108 | - **Interactive Commands**: Command CodeCrow directly from PR comments using `/ask`, `/analyze`, and `/summarize`. |
| 109 | +- **Issue Lifecycle**: Automatic tracking of resolved vs. open issues across analyses with deterministic and AI-based reconciliation. |
| 110 | +- **Bring Your Own Model**: Connect your preferred LLM provider — OpenRouter, Anthropic, Google, or OpenAI. |
26 | 111 |
|
27 | 112 | ## Documentation |
28 | 113 |
|
|
0 commit comments