Skip to content

ROADMAP.md

Latest commit

 

History

History
117 lines (81 loc) · 6.46 KB

ROADMAP.md

File metadata and controls

117 lines (81 loc) · 6.46 KB

Moss Roadmap

This is a living document. We update it as priorities shift based on community feedback and production learnings. If something here excites you, open an issue or PR — we move fast on contributions.

Shipped

  • Python SDK (moss) — async-first, type-safe
  • TypeScript SDK (@moss-dev/moss) — full feature parity with Python
  • Built-in embedding models (moss-minilm)
  • Hybrid search — combine semantic search with BM25 keyword matching
  • Custom embedding support (bring your own OpenAI, Cohere, etc.)
  • Metadata filtering ($eq, $and, $in, $near)
  • Document management (add, upsert, get, delete)
  • LangChain integration
  • DSPy integration
  • Pipecat voice agent integration
  • LiveKit voice agent integration
  • Next.js example app
  • VitePress search plugin
  • Docker deployment examples (ECS/K8s patterns)
  • WebAssembly runtime — client-side semantic search in the browser, no server required
  • Benchmarks directory — reproducible latency/throughput scripts comparing Moss vs Pinecone, Qdrant, and Chroma on standardized datasets
  • MCP server — expose Moss as a Model Context Protocol server so any MCP-compatible AI tool (Claude, Cursor, Windsurf) can do semantic search
  • Vercel AI SDK integration — retrieval provider for the Vercel AI SDK
  • Ollama + Moss + Pipecat reference architecture — an end-to-end local LLM voice agent: Ollama for LLM inference, Moss for retrieval, Pipecat for real-time audio. A single docker compose up to run the entire stack.

In Progress

  • CrewAI integration — Moss as a retrieval tool for CrewAI agents
  • Haystack integration — document store / retriever integration
  • Reranking support — plug in cross-encoder rerankers as a post-retrieval step

Next Up — Community Contributions Welcome

These are well-scoped and ready for contributors. Each one has (or will have) a corresponding GitHub issue with detailed instructions.

New SDK Bindings

Framework Integrations

  • AutoGen — retrieval-augmented tool for AutoGen agents
  • LlamaIndex — retriever and query engine integration
  • Semantic Kernel — .NET/Python retrieval plugin
  • LangGraph — retrieval node for stateful multi-agent workflows
  • Google ADK — Moss as a retrieval tool for Google's Agent Development Kit
  • OpenAI Agents SDK — Moss as a tool for the OpenAI agents framework
  • Smolagents — lightweight retrieval tool for Hugging Face's agent framework

Voice AI Ecosystem

  • Vapi integration — Moss retrieval tool for Vapi voice agents
  • Daily.co integration — real-time audio pipeline with semantic context injection
  • Twilio integration — retrieval for phone-based AI agents (IVR, call center bots)

Developer Tools

  • Moss CLI — manage indexes, run queries, import data, and inspect results from the terminal (moss index create, moss query, moss import)
  • VS Code extension — semantic search over your codebase directly from the editor sidebar

Search Quality

  • Multi-vector retrieval — support ColBERT-style late interaction models

Data Ingestion

  • Doc-parsing connectors — ingest PDF, DOCX, HTML, and Markdown files directly into Moss indexes
  • Chunking strategies — built-in text splitters (sentence, paragraph, recursive, semantic)
  • Web crawling — crawl a URL and index the content

Future

These are bigger bets we're exploring. They're directional, not committed — community input will shape what gets built.

Local-First AI Stack

  • vLLM-based local inference + local search — a fully local pipeline: your model, your embeddings, your search, your hardware. No API calls. This is a natural fit for the privacy-first voice AI use case and can meaningfully cut latency for on-premise deployments.

Evaluation & Quality

  • LLM-as-a-judge evaluation framework — automated retrieval quality scoring using LLM judges. We want to lay the foundation and let the community decide the direction — what metrics matter, which judges to support, how to benchmark fairly.
  • Retrieval quality dashboard — visualize query performance, relevance scores, and failure modes over time

Browser & Edge

  • Edge runtime support — run Moss in Cloudflare Workers, Deno Deploy, and Vercel Edge Functions

Advanced Retrieval

  • Query expansion — LLM-powered query rewriting to improve recall on short or ambiguous queries
  • Sparse-dense fusion (SPLADE) — learned sparse retrieval to complement BM25 hybrid, improving precision on rare terms
  • Contextual retrieval — pre-chunking contextualization to make every chunk self-contained and more retrievable

Data Connectors

Connect knowledge sources to Moss without writing custom ETL.

  • GitHub connector — index code, issues, PRs, and docs from repositories
  • Notion connector — sync and index Notion workspace pages
  • Confluence connector — enterprise knowledge base indexing
  • S3/GCS sync — auto-index documents from cloud storage buckets on upload

How to Contribute

  1. Pick something from "Next Up" — these are ready for PRs
  2. Check the issues — look for good first issue and help wanted labels
  3. Propose something new — open an issue describing what you want to build. We're open to ideas that aren't on this list.
  4. Read the Contributing Guide — fork, branch from main, PR

If you're unsure where to start, drop a message in Discord and we'll point you in the right direction.