Skip to content

Roadmap

Jump to bottom
Sanni heruwala edited this page Jun 13, 2026 · 2 revisions

Roadmap

This page mirrors docs/roadmap.md and tags each item with the current status. Open issues + design docs are linked inline so contributors can pick things up cleanly.

Phase 1. MVP — ✅ shipped

  • ✅ Trino HTTPS connector
  • ✅ Notebook UI (SQL, Markdown, AI prompt cells)
  • ✅ Result table + profile + chart + AI summary tabs
  • ✅ SQL guard with read-only default
  • ✅ Mock AI + OpenAI / Anthropic / Ollama options
  • ✅ Internal Knowledge Notebook + infographic HTML export
  • ✅ Notebook persistence (local JSON)
  • ✅ FastAPI HTTP layer + Typer CLI + Docker

Phase 2. Polish & analyst tooling — ✅ shipped

  • ✅ AI chat over knowledge sources (multi-turn Ask AI)
  • ✅ Saved connections store + per-instance credentials
  • ✅ Per-user namespacing (auth, GitHub OAuth, API tokens, admin invites)
  • ✅ AG Grid for power-user result navigation
  • ✅ Knowledge layer as a first-class slide-in panel with full CRUD
  • ✅ Server-side query cancellation (DuckDB / Trino / Postgres / Redshift / MySQL / MariaDB) — Stop button now reaches the engine
  • ✅ Richer profiling — per-column histograms + normalised-MI Related columns
  • ✅ Headless-browser PDF / PNG infographic export via Playwright (in the [exports] extra)
  • ✅ Summarize result button (in-cell, replaces Explain SQL)
  • ✅ Unified connection control in the left sidebar
  • ✅ NotebookLM-style knowledge: [n] citation chips + Studio (Overview / FAQ / Study guide / Ask next)
  • ✅ Onboarding tour + Quick-connect templates (Supabase / Neon / Vercel / Railway / Render / Heroku / DuckDB) + "Start here" empty state (v0.7.21)
  • ✅ Bundled sample notebook + first-run seeder (v0.7.21)
  • ✅ Page refresh preserves cell results, history dialog redesign, Plotly charts in published HTML (v0.7.20)
  • ✅ HD chart renderer (SVG-crisp), virtual "files" catalog in metadata (v0.7.19)
  • ✅ SQL formatter button per cell (v0.7.17, shipped in v0.7.18 image)
  • ❌ Cursor AI provider — removed in v0.7.20 (no public chat-completions API exists)
  • 🟡 Python cell type — deferred. RedNotebook AI stays SQL-first.

Phase 3. More connectors — ✅ mostly shipped

  • ✅ DuckDB
  • ✅ PostgreSQL, MySQL, MariaDB, SQLite, MSSQL, Snowflake, BigQuery, Redshift, Oracle, ClickHouse, Databricks (11 SQL connectors via a generic SQLAlchemy backend)
  • ✅ Drivers bundled by default — no pip install ...[extras] step needed
  • ✅ Drag-and-drop file uploads — CSV / TSV / Parquet / JSON / JSONL / NDJSON, auto-registered as DuckDB views
  • 🟡 Athena connector
  • 🟡 Excel (.xlsx) — needs DuckDB's excel extension autoloaded
  • 🟡 Google Sheets upload (requires OAuth)

Phase 4. Collaboration / SaaS — 🟡 in progress

  • ✅ Rate limiting via slowapi on auth + AI routes
  • ✅ Audit log (rednotebook/audit/log.py) — surfaced in the admin UI
  • ✅ Git-backed notebooks — per-user notebook dir is a git repo with autosave commits + a History dialog that restores any version
  • ✅ One-click static HTML share links — POST /api/notebooks/{id}/publish mints a token, page served unauthenticated at /published/{token}
  • 🟡 Live sharing + threaded comments on the live notebook (static publish covers read-only sharing today)
  • 🟡 Scheduled queries / alerts
  • 🟡 Dashboard publishing — notebook → parameterised app with input widgets (the publish HTML is the foundation)
  • 🟡 dbt + Airflow integrations
  • 🟡 Semantic layer
  • 🟡 Cross-source joins via DuckDB ATTACH — design doc ready
  • 🟡 Reactive notebooks (Marimo-style auto-rerun) — design doc ready
  • 🟡 Full RBAC / SSO (partial — GitHub OAuth + admin invites shipped)
  • 🟡 Hosted SaaS version

Want to pick something up?

Two well-scoped design docs are sitting in docs/design/ with tracking issues #3 and #4 above. Both have the full backend shape, frontend shape, edge cases, and acceptance criteria spelled out.

For smaller starters see Contributing.

Clone this wiki locally