Repowise
Repowise is an open-source, self-hostable codebase-intelligence platform for AI coding agents and engineering teams. It builds one living index of source code, dependencies, Git history, tests, documentation and architectural decisions so agents can answer questions with evidence and reviewers can assess change risk before merging.
At a Glance
- Individual developers and open-source maintainers
- Software engineering teams and engineering organizations
- Organizations with large, legacy or multi-repository codebases
- Teams adopting AI coding agents
- +2 more
AI Tools by Repowise
(1)
Repowise
Codebase Intelligence for Agents
Discussions
No discussions yet
Be the first to start a discussion about Repowise
Latest News
Repowise v0.60.0: PR bot learns from team review decisions; merge-decision PR pages, C4 knowledge graph and cited agent answers added.
Repowise v0.59.0: Salesforce and COBOL mapping, cross-service System Map and repository decision recording.
Repowise v0.58.0: action-oriented dashboard pages, zoomable Code Health folder map, guided CLI init and CI gates.
Repowise v0.57.0: evidence checklists in PR-bot comments, Change Records and policy/evidence packs.
Products & Services
AGPL-3.0, self-hostable software installable with pip/uv/pipx. It indexes code, Git history, tests and decisions locally, then exposes a dashboard, CLI commands and an MCP server without requiring an API key for indexing.
Managed repository indexing with public repository pages, generated documentation, semantic search, codebase chat, persistent dashboards, private-repository sync and hosted MCP access.
GitHub pull-request analysis that reports blast radius, callers outside the PR, ownership, risky files, tests, companion files and change risk; Pro adds AI review and decision learning.
An announced/early hosted capability described as models trained on real commits and fixes to rank files a change may touch, identify omissions and explain risk; the source presents it as coming to repowise.dev.
Market Position
Repowise positions itself as an open-source, local-first codebase-intelligence layer rather than only code search or documentation. Its official comparisons place it against Sourcegraph, CodeScene, SonarQube, Semgrep, Greptile, DeepWiki, Google Code Wiki and CodeSee: Repowise differentiates with a unified graph/history/health/decision index, MCP access for coding agents, deterministic and self-hostable analysis, Git-based change risk and a hosted option for private repositories.
Leadership
Founders
Raghav Chamadiya
AI-tools engineer and co-founder of Repowise; he designed the deterministic-first analysis and authored the defect-prediction validation study behind Repowise's code-health score. His public profile also describes him as building AI tools for engineers.
Swati Ahuja
Co-founder of Repowise and co-founder of thita.ai, an AI technical-interview practice product. Repowise's GitHub Sponsors profile says she previously worked at Salesforce and Intuit.
Executive Team
Raghav Chamadiya
Co-founder
AI-tools engineer; author of Repowise's defect-prediction validation study and designer of its deterministic code-health analysis.
Swati Ahuja
Co-founder
Co-founder of Repowise and thita.ai; previously worked at Salesforce and Intuit.
Founding Story
The founders started Repowise to address the repeated cost and unreliability of AI coding agents reading repositories file by file, without architectural context, history, tests or memory of prior decisions. Their initial vision was an open-source, local-first intelligence layer that indexes a repository once and lets agents query that index instead of repeatedly grepping through code; the project was released openly by Raghav Chamadiya and Swati Ahuja in 2026.
Business Model
Revenue Model
Freemium hosted SaaS: public repositories and the open-source local product are free; paid hosted subscriptions monetize private-repository indexing, automatic synchronization, security scanning, hosted MCP/chat, team workspaces and PR-bot/AI usage. AI work is metered through prepaid or included credits. Enterprise self-hosting is offered through custom agreements.
Pricing Tiers
Two public repositories, 10 AI answers per month, public repository page and documentation, semantic search, deterministic analysis, PR bot on public repositories and up to 250 MB public indexing.
Private repositories through the GitHub App, push-triggered auto-sync, full history, security scanning, hosted Claude.ai/ChatGPT access, codebase chat, AI PR review and $5/month AI credits.
Shared repositories and credit pool, organization GitHub App, roles, team-wide PR bot, portfolio health, cross-repository ownership/workspaces and concurrent indexing.
Self-hosted/offline enterprise deployment and scoped requirements such as SSO, SCIM and private-network operation.
Target Markets
- Individual developers and open-source maintainers
- Software engineering teams and engineering organizations
- Organizations with large, legacy or multi-repository codebases
- Teams adopting AI coding agents
- Security, platform-engineering and developer-productivity teams
- Enterprises requiring self-hosted or air-gapped analysis
- Giving Claude Code, Codex, Cursor, Copilot and other MCP-capable agents repository context before they edit
- Understanding unfamiliar or inherited codebases
- Predicting pull-request blast radius and deciding which tests to run
- Code health, technical-debt prioritization, refactoring planning and dead-code cleanup
- Keeping generated documentation and architectural decisions synchronized with code
- Security and supply-chain checks in CI, including secrets, CVEs and SBOM/VEX workflows
- Developers at Microsoft, Red Hat, Siemens, AWS, Atlassian, Accenture, NVIDIA, SAP, Meta, Thoughtworks, Wix, EY, Deutsche Telekom and other companies shown on Repowise's official site