# Moss

> Moss is a real-time semantic search runtime for AI agents, voice agents, copilots, and multimodal apps. It delivers sub-10ms lookups with zero infrastructure. Built in Rust and WebAssembly. Backed by Y Combinator. Built by InferEdge Inc.

## What Moss Does

Moss runs semantic search inside your agent runtime — browser, edge, on-device, or cloud — so retrieval is local, fast, and private. Connect your data once; Moss indexes, distributes, and keeps indexes fresh. No database clusters, no replicas, no DevOps.

## SDKs

- JavaScript/TypeScript: npm install @inferedge/moss
- Python: pip install inferedge-moss
- Package name (JS): @inferedge/moss
- Package name (Python): inferedge-moss

## Authentication

Moss uses project-based credentials. Get your MOSS_PROJECT_ID and MOSS_PROJECT_KEY from the Moss Portal at https://portal.usemoss.dev. Set them as environment variables.

## Embedding Models

- moss-minilm (default): Fast, lightweight, ideal for edge/offline use
- moss-mediumlm: Higher accuracy with reasonable performance
- Bring Your Own Embeddings (BYOE): Pass pre-computed embeddings via DocumentInfo.embedding at index time and QueryOptions.embedding at query time. Use any external model (OpenAI, Cohere, local, etc.). See python/custom_embedding_sample.py.

## Key Concepts

- Index: A searchable structure built from your documents. Supports multiple indexes per project.
- Document: Has id (string), text (string), and optional metadata (string map). Upserts replace matching ids.
- Retrieval: Supports vector similarity (semantic), keyword/BM25, and hybrid search. Controlled via alpha parameter (1.0 = pure semantic, 0.0 = pure keyword).
- Storage: Indexes stored on-device by default. Optional cloud sync for backup and sharing.
- Client Lifecycle: Create index -> Load index -> Query -> Upsert/Delete docs -> Delete index when done.

## Getting Started

- [Quickstart Guide](https://docs.moss.dev/docs/start/quickstart): Install SDK, configure credentials, create and query an index in minutes
- [What is Moss?](https://docs.moss.dev/docs/start/what-is-moss): Overview and architecture
- [Core Concepts](https://docs.moss.dev/docs/start/core-concepts): Index, Document, Embeddings, Retrieval, Storage explained

## Documentation

- [SDK Overview](https://docs.moss.dev/docs/reference/sdk): High-level SDK guide with examples in JS and Python
- [JS SDK Reference](https://docs.moss.dev/docs/reference/js/api): Full TypeScript API — MossClient class, interfaces, types
- [MossClient API (JS)](https://docs.moss.dev/docs/reference/js/classes/MossClient): Complete method reference — createIndex, loadIndex, query, addDocs, deleteDocs, getDocs, getIndex, listIndexes, deleteIndex, getJobStatus
- [API Reference (REST)](https://docs.moss.dev/docs/api-reference/v1/getting-started/introduction): Control Plane API at https://service.usemoss.dev/v1

## Integration Guides

- [Authentication](https://docs.moss.dev/docs/integrate/authentication): Setting up project credentials
- [Indexing Data](https://docs.moss.dev/docs/integrate/indexing-data): Creating indexes, upserting documents, choosing models, chunking tips
- [Retrieval](https://docs.moss.dev/docs/integrate/retrieval): Semantic, keyword, and hybrid search with alpha tuning and metadata filters
- [Storage & Persistence](https://docs.moss.dev/docs/integrate/storage-persistence): Local storage, cloud sync, backup strategies
- [Deployment / Production](https://docs.moss.dev/docs/integrate/deployment-production): Production checklist, security, observability

## Use Cases

- [Sub-10ms Knowledge Retrieval](https://docs.moss.dev/docs/build/offline-first-search): Offline-first search for knowledge bases
- [Voice Agent with LiveKit](https://docs.moss.dev/docs/integrations/livekit): Build a voice AI agent with Moss + LiveKit + OpenAI + Deepgram

## Pricing

- Developer (Free): $5/month free credits, build & index datasets, unlimited local queries, shared infra
- Hobbyist ($30/mo): Continuous sync engine, unlimited projects & indexes, session replays (7 days), file uploads
- Start-up ($200/mo): Hot path cloud search, 150 concurrent sessions, session replays (30 days), priority ingest, email support
- Enterprise (Custom): Custom scale, white glove onboarding, 99.9% SLA, SSO, 24/7 Slack support, SOC2 & HIPAA compliance
- Details: https://www.moss.dev/pricing

## Sample Code

- GitHub: https://github.com/usemoss/moss-samples
- JS samples: javascript/comprehensive_sample.ts, javascript/load_and_query_sample.ts
- Python samples: python/comprehensive_sample.py, python/load_and_query_sample.py, python/custom_embedding_sample.py (BYOE)

## Links

- [Website](https://www.moss.dev)
- [Documentation](https://docs.moss.dev/docs)
- [Moss Portal](https://portal.usemoss.dev)
- [GitHub](https://github.com/usemoss)
- [Discord](https://moss.link/discord)
- [Release Notes](https://docs.moss.dev/docs/changelog)

## Legal

- [Privacy Policy](https://docs.moss.dev/docs/privacy)
- [Terms of Service](https://docs.moss.dev/docs/tos)

## Company

Moss is built by InferEdge Inc., based in San Francisco, CA. Backed by Y Combinator. Trusted by 500+ teams including Microsoft, EPAM, UC Berkeley, Carnegie Mellon, Podium, and Stanford.