Working examples for Moss — the real-time search runtime for AI agents.
Each sample is kept thin on purpose. Copy what you need straight into your own project, swap in your data, and go. Python, TypeScript, Next.js, and Pipecat voice agents are all covered.
Questions? Join our Discord.
- Sign up at moss.dev and create an index from the dashboard.
- Open View secrets and save the values to your
.env:
MOSS_PROJECT_ID=your_project_id
MOSS_PROJECT_KEY=your_project_key
For full setup details and pricing, see the docs.
Install deps: pip install -r python/requirements.txt, then run any script with python path/to/sample.py.
comprehensive_sample.py— end-to-end flow: session creation, context building, streaming responsesload_and_query_sample.py— ingest domain knowledge, then querycustom_embedding_sample.py— create an index with custom OpenAI embeddings and run queries
Install deps: npm install inside javascript/, then run with npm run start -- path/to/sample.ts.
comprehensive_sample.ts— full workflow in TypeScript, ready for Nodeload_and_query_sample.ts— index FAQs and issue targeted queriescustom_embedding_sample.ts— provision an index, push OpenAI embeddings, query
A web-based semantic search interface using Next.js 15 and Server Actions. Shows how to call Moss securely from the server while serving a responsive UI.
cd next-js && npm install && npm run devOpen http://localhost:3000. See the next-js/ directory for details.
A voice bot that plugs Moss retrieval into Pipecat's real-time pipeline — sub-10ms semantic search for a customer support agent.
pipecat-moss/pipecat-quickstart/— ingest FAQs withcreate-index.py, wire into Pipecat, deploy to Pipecat Cloud- For deeper context, see the Pipecat-Moss repo
- Documentation — API reference, guides, architecture
- Website — product overview and pricing
- Launch YC post
If you spot gaps or want another language example, open an issue or PR. We track feedback closely.
See our Contributing Guide for details.