Features Overview

Lightfast is memory built for teams. It helps you search everything your team knows, understand context across your entire codebase, and trace decisions from idea to implementation.

Core Capabilities

Intelligent Search

Search by meaning, not keywords. Lightfast understands what you're asking for and finds relevant content even when it uses different terminology.

What you can do:

• Find answers to technical questions across all your documentation
• Discover related implementations and design patterns
• Trace decisions from RFCs through PRs to production

Memory Organization

Your team's knowledge is structured into chunks, observations, summaries, and profiles—each optimized for different types of retrieval.

What you can do:

• Search long documents and get precise sections, not entire files
• Capture atomic moments like decisions, incidents, and highlights
• Access aggregated summaries by person, repo, or time period
• Find experts based on their contributions and expertise

Graph Relationships

Lightfast tracks connections between people, code, and decisions to answer "who owns what" and "what depends on what" questions.

What you can do:

• Identify code owners and subject matter experts
• Trace dependencies between services and libraries
• Connect RFCs to implementing PRs and related discussions
• Find collaboration patterns across your team

Citations & Sources

Every answer includes verifiable sources. No claims without evidence, no summaries without citations.

What you can do:

• Verify every claim with links to source documents
• See relevance scores and timestamps for each citation
• Understand why specific sources were selected
• Trust that information comes from actual content, not hallucinations

Quality & Evaluation

Lightfast continuously measures search quality and adapts to your team's usage patterns through calibration and feedback loops.

What you can do:

• Benefit from automatic ranking improvements over time
• See which metrics drive search quality (recall, precision, MRR)
• Influence results through your usage patterns
• Trust that quality is measured and optimized continuously

Privacy & Security

Complete tenant isolation, permission-aware search results, and full audit trails ensure your data stays secure.

What you can do:

• Search with confidence that your data never leaks to other workspaces
• Rely on GitHub permissions to control who sees what
• Access audit logs for compliance and security reviews
• Meet SOC 2 Type II requirements for enterprise deployment

How Features Work Together

These features combine to create a complete team memory system:

1. Indexing: Your repositories, PRs, issues, and docs are captured and organized
2. Search: Semantic understanding finds relevant content across all indexed material
3. Relationships: Graph connections surface ownership and dependencies
4. Citations: Every answer links back to verifiable sources
5. Quality: Continuous evaluation and calibration improve results over time
6. Security: Permissions and isolation keep everything safe and private

Next Steps

Ready to dive deeper? Explore individual features or jump into the API:

• Search & Retrieval — How semantic search works
• Memory Organization — Chunks, observations, summaries, profiles
• Architecture — Technical deep dive
• API Reference — Four simple endpoints
• Examples — Real-world implementations