Deepnote
Free tierThe collaborative data workspace built for humans and AI agents
Free tier available·All audiences·API available·Open source
Key strengths
AI-powered notebook with code generation, explanation, and debuggingReal-time collaboration with commenting, versioning, and sharingUnified Python, SQL, and R environment in a single IDEBeautiful data apps, dashboards, and publishable reports from notebooksEnterprise-grade security: SOC 2, HIPAA, GDPR, RBAC, and SSO
Free tier + paid plans
San Francisco, USA
Founded 2019
Self-hostable
No ratings yet
Developer & Technical Documentation
Notebook Environment
Deepnote is now open-source and supports Python 3.x, SQL (with Jinja templating via dbt), and R. Projects run in isolated cloud containers with configurable hardware — from standard CPUs to high-memory GPU instances (for model training and ML workloads).
AI & Agent Capabilities
- Deepnote AI is context-aware: it reads your schema, variable state, and prior cells to generate, complete, refactor, explain, and debug code.
- Autonomous data agents can be built by chaining notebook "skills" — reusable notebook templates callable via
deepnote.run(skill="...", period="..."). - Agents can be launched from external tools like Claude Code or Cursor using the Deepnote CLI and REST API.
API & Integrations
- REST API available at
deepnote.com/api/v1/— supports serving models directly from notebooks as live endpoints. - VS Code extension allows starting locally and continuing in the cloud, publishing figures, running cells as skills, and scheduling notebooks.
- Supports Spark, Snowpark, and terabyte-scale data via serverless cluster connections.
Security & Deployment
- Compliant with SOC 2, HIPAA, GDPR.
- Supports RBAC, SAML/OIDC SSO, and directory sync.
- Custom managed deployment options available for Enterprise customers.
- Notebooks can be deployed as scheduled jobs (hourly/daily/weekly/monthly) or API endpoints.