# Quark

## Docs

- [Authentication](https://mintlify.wiki/drona-gyawali/Quark/api-reference/authentication.md): How to authenticate requests to the Quark API using Supabase JWTs.
- [Chat](https://mintlify.wiki/drona-gyawali/Quark/api-reference/endpoints/chat.md): Query documents, retrieve paginated chat history, and delete conversations.
- [Ingest document](https://mintlify.wiki/drona-gyawali/Quark/api-reference/endpoints/ingest.md): Process an uploaded document into chunks and store its embeddings in the vector store.
- [Sessions](https://mintlify.wiki/drona-gyawali/Quark/api-reference/endpoints/sessions.md): Create, list, rename, and delete chat sessions.
- [Get upload URL](https://mintlify.wiki/drona-gyawali/Quark/api-reference/endpoints/upload.md): Request a presigned S3 URL to upload a document directly to object storage.
- [API overview](https://mintlify.wiki/drona-gyawali/Quark/api-reference/overview.md): Base URLs, authentication requirements, and a quick reference for all Quark API endpoints.
- [Ingestion pipeline](https://mintlify.wiki/drona-gyawali/Quark/architecture/ingestion-pipeline.md): A detailed walkthrough of how Quark transforms a raw document into searchable, vision-enriched vector embeddings.
- [Memory system](https://mintlify.wiki/drona-gyawali/Quark/architecture/memory-system.md): How Quark's dual-layer memory architecture combines Redis short-term memory and Mem0 long-term memory to maintain persistent, session-aware context.
- [Architecture overview](https://mintlify.wiki/drona-gyawali/Quark/architecture/overview.md): A high-level look at Quark's modular RAG architecture — how documents flow from ingestion to retrieval.
- [Vector search and re-ranking](https://mintlify.wiki/drona-gyawali/Quark/architecture/vector-search.md): How Quark retrieves relevant document chunks using a two-stage pipeline: Qdrant vector similarity search followed by VoyageAI re-ranking.
- [Configuration](https://mintlify.wiki/drona-gyawali/Quark/configuration.md): Complete reference for all environment variables and pipeline constants that control Quark's behavior.
- [CLI usage](https://mintlify.wiki/drona-gyawali/Quark/guides/cli-usage.md): Start the Quark CLI, register a username, and use the full command set to ingest documents, query your knowledge base, and manage sessions.
- [Document ingestion](https://mintlify.wiki/drona-gyawali/Quark/guides/document-ingestion.md): Ingest PDF documents into Quark's vector store via the CLI or the REST API. Each document is chunked, embedded, and tagged with optional metadata.
- [Querying documents](https://mintlify.wiki/drona-gyawali/Quark/guides/querying-documents.md): Ask questions against your ingested documents using the CLI or the REST API. Quark retrieves the most relevant chunks and returns a grounded answer.
- [Session management](https://mintlify.wiki/drona-gyawali/Quark/guides/session-management.md): Sessions are named conversation threads with isolated chat history. Create, switch, rename, and delete sessions from the CLI or the API.
- [Introduction](https://mintlify.wiki/drona-gyawali/Quark/introduction.md): Quark is a high-performance RAG system that lets you chat with your documents — grounded in what you uploaded, never what the model imagines.
- [Quickstart](https://mintlify.wiki/drona-gyawali/Quark/quickstart.md): Clone Quark, configure your services, and chat with your first document in under 10 minutes.
- [Running the API server](https://mintlify.wiki/drona-gyawali/Quark/self-hosting/api-server.md): Start the Quark API server, understand its base path, authentication model, and health check.
- [Environment variables](https://mintlify.wiki/drona-gyawali/Quark/self-hosting/environment-variables.md): Reference for every variable in .env.example, grouped by service.
- [Infrastructure](https://mintlify.wiki/drona-gyawali/Quark/self-hosting/infrastructure.md): System requirements and setup steps for self-hosting Quark.

## OpenAPI Specs

- [openapi](https://mintlify.wiki/drona-gyawali/Quark/api-reference/openapi.json)