v1.0.0: The AI-Native Internal Developer Portal

🎉 Corpus v1.0.0 - The AI-Native Internal Developer Portal

We are thrilled to announce the official 1.0 release of Corpus! This release transforms Corpus from a simple script into an enterprise-ready, zero-database Model Context Protocol (MCP) server that empowers your AI agents to holistically understand your entire organization.

✨ Key Features

• Multi-Provider Architecture: Native, out-of-the-box support for GitHub, GitLab, Bitbucket, and Azure DevOps.
• Local Semantic Search: Optionally enable state-of-the-art vector embedding search powered locally by HuggingFace's Transformers.js. Zero external API calls, zero latency, and $0 cost.
• Backstage System Graph: Automatically parses catalog-info.yaml files to generate an active, bidirectional relationship graph (ownedBy, dependsOn, providesApi) so AI agents can traverse your service boundaries.
• Automated Skill Directory: Dynamically scans all repositories for .[client]/skills/**/SKILL.md (supporting .agents, .claude, .cursor, etc.) to extract YAML frontmatter, allowing agents to instantly learn your standard operating procedures.
• Global Code & Issue Search: Proxies to your VCS provider's native search APIs so agents can find Code, PRs, and Discussions globally across the org.
• API Schema Aggregation: Automatically indexes OpenAPI/Swagger schemas.
• Doc Gap Reporting: Agents can dynamically file missing documentation as tracking items in a central repository.
• Zero-Database & Dockerized: Runs anywhere instantly via npm start or Docker. No Postgres, no Pinecone, no infrastructure bloat.

Thank you to all contributors who made V1 a reality!