# smarts.bio

> AI-powered bioinformatics platform. Run BLAST, GATK, variant calling, RNA-seq, and multi-step pipelines on your own data through a REST API, Python/TypeScript SDKs, a cross-platform CLI, and MCP servers.

Developers authenticate with `sk_live_` API keys and call the gateway at https://api.smarts.bio/v1. A machine-readable OpenAPI 3.1 spec is published at https://smarts.bio/openapi.json. AI agents can connect directly over the Model Context Protocol at https://mcp.smarts.bio (Streamable HTTP) — for agentic use, prefer MCP over scraping these docs.

## Getting Started
- [Quick Start](https://smarts.bio/docs/getting-started): Call the smarts.bio bioinformatics agent from Python, TypeScript, or any HTTP client.
- [Authentication](https://smarts.bio/docs/authentication): Authenticate requests with sk_live_ API keys; scopes and header formats.
- [Available Tools](https://smarts.bio/docs/getting-started/tools): Catalog of bioinformatics tools you can run directly (BLAST, GATK, and more).
- [Available Databases](https://smarts.bio/docs/getting-started/databases): Biological databases the agent can search (NCBI, STRING, PDB, KEGG, ClinVar, patents).
- [Predefined Pipelines](https://smarts.bio/docs/getting-started/pipelines): Ready-made multi-step pipelines (QC, WES alignment, RNA-seq) and how to run them.

## CLI & MCP
- [CLI](https://smarts.bio/docs/cli): Install and use the cross-platform `smarts` CLI for terminals, scripts, and coding agents.
- [MCP Overview](https://smarts.bio/docs/mcp): Connect AI agents to smarts.bio over the Model Context Protocol.
- [Hosted MCP](https://smarts.bio/docs/mcp/hosted): Connect ChatGPT, Gemini, or Claude-web to the hosted MCP server at mcp.smarts.bio (Streamable HTTP).
- [Local MCP](https://smarts.bio/docs/mcp/local): Run a local stdio MCP server for Claude Desktop, Cursor, and Claude Code.

## API Reference
- [Workspaces](https://smarts.bio/docs/reference/workspaces): List and select workspaces that scope files, pipelines, and queries.
- [Query](https://smarts.bio/docs/reference/query): POST /v1/query and /v1/query/stream — run the AI agent synchronously or over SSE.
- [Tools](https://smarts.bio/docs/reference/tools): List tools and run a specific tool directly via POST /v1/tools/:id/run.
- [Pipelines](https://smarts.bio/docs/reference/pipelines): Create, list, poll, and cancel asynchronous pipeline runs.
- [Visualizations](https://smarts.bio/docs/reference/visualizations): Generate viewer links and render plots (volcano, heatmap, PCA) from workspace files.
- [Conversations](https://smarts.bio/docs/reference/conversations): List conversations and fetch full message history.
- [Files](https://smarts.bio/docs/reference/files): Upload, list, download, move, rename, and delete workspace files.

## Databases
- [SmartsMatch ✦](https://smarts.bio/docs/databases/smarts-match): smarts.bio proprietary matching database.
- [Biograph ✦](https://smarts.bio/docs/databases/biograph): smarts.bio knowledge graph of genes, proteins, variants, diseases, pathways, and drugs.
- [BLAST / Sequence Search](https://smarts.bio/docs/databases/blast): Nucleotide and protein sequence similarity search.
- [NCBI](https://smarts.bio/docs/databases/ncbi): PubMed, GenBank, and Gene access via the NCBI toolkit.
- [STRING (PPI)](https://smarts.bio/docs/databases/string): Protein–protein interaction networks from the STRING database.
- [PDB / AlphaFold](https://smarts.bio/docs/databases/pdb): Experimental and predicted protein structures.
- [ClinVar & Variants](https://smarts.bio/docs/databases/clinvar): Clinical variant interpretation and annotation.
- [KEGG & Reactome](https://smarts.bio/docs/databases/kegg): Pathway and reaction databases.
- [Patents](https://smarts.bio/docs/databases/patents): Global patent search via Google Patents Public Data for IP and prior-art research.

## Examples
- [Genomics](https://smarts.bio/docs/examples/genomics): End-to-end genomics workflows (variant calling, alignment).
- [Transcriptomics](https://smarts.bio/docs/examples/transcriptomics): RNA-seq and expression analysis examples.
- [Proteomics](https://smarts.bio/docs/examples/proteomics): Protein analysis, domains, and interaction examples.
- [Structural Biology](https://smarts.bio/docs/examples/structural-biology): Structure retrieval and analysis examples.
- [Metabolomics](https://smarts.bio/docs/examples/metabolomics): Metabolite and pathway analysis examples.
- [Literature & Databases](https://smarts.bio/docs/examples/literature-databases): Literature search and database lookup examples.
- [Pipelines](https://smarts.bio/docs/examples/pipelines): Worked examples of running multi-step pipelines.

## SDKs
- [Python SDK](https://smarts.bio/docs/sdks/python): The `smartsbio` PyPI package for Python scripts and Jupyter notebooks.
- [TypeScript SDK](https://smarts.bio/docs/sdks/typescript): The `@smartsbio/sdk` npm package for Node.js and TypeScript.

## Reference
- [Error Codes](https://smarts.bio/docs/errors): API error response format and the full list of error codes.

## Resources
- [OpenAPI specification](https://smarts.bio/openapi.json): Machine-readable OpenAPI 3.1 description of the public REST API.
- [Hosted MCP server](https://mcp.smarts.bio): Streamable HTTP MCP endpoint for AI agents (ChatGPT, Gemini, Claude).
- [Python SDK on PyPI](https://pypi.org/project/smartsbio/): `pip install smartsbio`.
- [TypeScript SDK on npm](https://www.npmjs.com/package/@smartsbio/sdk): `npm install @smartsbio/sdk`.