Developer Documentation Search

Part of the worlds-biggest-software-project initiative.

Semantic search across all internal and external developer documentation, returning cited answers rather than link lists.

Developer Documentation Search is an AI-native, open-source search layer for engineering teams. It unifies internal docs, API references, code, and external knowledge into a single, permission-aware index, designed for DX engineers, DevRel teams, and platform owners frustrated by the cost or quality of incumbent search tools.

Why Developer Documentation Search?

Algolia gets expensive fast. Pricing of $1 per 1K searches beyond the free tier means mid-market teams consistently overpay as their docs portals grow.
Enterprise search is overkill and opaque. Glean ($7.2B valuation, $208M ARR) and Coveo target $100K+/yr enterprise contracts; smaller teams need a focused doc-search tool, not a full enterprise knowledge platform.
Confluence search is a recurring frustration. Engineering managers cite poor in-tool search quality as a productivity blocker, with the global enterprise search market projected to reach USD 8.9 billion by 2026.
RAG-based answers are now expected UX. Link-list results no longer meet developer expectations; semantic search and grounded answer generation are becoming table stakes.
Permission-aware search is still a gap. Respecting source-system ACLs (OAuth 2.0 / OIDC) is a known blocker for enterprise adoption that not all incumbents handle cleanly.

Key Features

Core Search

Full-text search across internal and external documentation
Markdown and HTML parsing and indexing
Sub-100ms search response targets
Faceted filtering by product, language, and version
Real-time indexing updates

API & Code Awareness

REST API documentation support
API schema integration with OpenAPI and GraphQL
Code example extraction and ranking
Versioned documentation with fallback behaviour

AI & Semantic Layer

Semantic search that understands developer intent
NLP-based query expansion and refinement suggestions
Natural language question answering over the corpus
Automatic documentation generation from code comments

Platform & Operations

Integration with major doc platforms (Markdown, GitBook) and internal sources
Mobile-responsive search interface
User management and search analytics
Popular-query reporting for content teams

AI-Native Advantage

Unlike incumbents that bolt AI onto a keyword index, this project is designed AI-first: heterogeneous sources (Confluence, GitHub, Notion, OpenAPI specs, Slack threads) are ingested into a single ranked answer surface with cited sources and page previews. Embeddings refresh automatically on commit, eliminating manual re-indexing. The system also detects documentation coverage gaps by identifying concepts developers search for but no page answers adequately, and offers a conversational "ask the docs" assistant that reformulates incomplete queries before retrieving results.

Tech Stack & Deployment

Expected to support both self-hosted and managed deployment, in line with open-source incumbents like Typesense, Meilisearch, and OpenSearch. Relevant standards include OpenAPI/Swagger for API corpora, OpenSearch as a deployment target for enterprise internal search, RSS/Atom for changelog and blog ingestion, OAuth 2.0 / OIDC for permission-aware results, and pluggable vector embedding providers (OpenAI, Cohere, Voyage) given that embedding choice materially affects recall.

Market Context

The global enterprise search market is projected to reach USD 8.9 billion by 2026, with over 70% of organisations citing it as critical for employee productivity. Pricing across incumbents ranges from free self-hosted (Typesense, Meilisearch) through usage-metered SaaS (Algolia at $1/1K searches) to enterprise contracts (Glean, Coveo at $100K+/yr). Primary buyers are DX engineers managing internal docs portals, DevRel teams publishing public API references, engineering managers dissatisfied with Confluence search, and platform teams owning the internal knowledge graph.

Project Status

This project is in the research and specification phase.
Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.