Bi-Dash

A business intelligence workspace with local AI, built on Redash.

Bi-Dash User Guide → — available features, improvements over upstream Redash, practical examples, roles, AI settings and troubleshooting.

• Query connected data sources and build interactive dashboards.
• Draft SQL with an integrated AI assistant powered by local Ollama.
• Create saved draft dashboards from prompts through AI Assistant → Build dashboard, including queries, charts and initial results. See AI_DASHBOARDS.md.
• Choose Ollama, OpenAI, Anthropic, or an OpenAI-compatible service in Settings → AI Configuration.
• Save shared schemas and business definitions in Settings → AI Context.
• Manage Admin, Editor and Reader roles, export permissions and query limits in Settings → Roles & Access. See ACCESS_CONTROL.md.
• A custom rising-signal logo with blue and yellow accents and anthracite navigation.
• Local Docker deployment with persistent PostgreSQL, Redis, and Ollama storage.

Portable Linux server installation

For a public server with a domain and automatic HTTPS, follow Public server deployment. Visitors access the site in their browser without an SSH tunnel.

For automatic secrets and a repeatable private server installation, see Linux deployment. Run bash scripts/linux.sh install after installing Docker.

Install from scratch: Windows, Linux and macOS

Follow the complete installation guide for prerequisites, first setup, AI configuration, updates and troubleshooting:

• Windows installation — Docker Desktop, Git and PowerShell 7.
• Linux installation — Docker Engine, Compose v2, Git and OpenSSL.
• macOS installation — Docker Desktop for Intel or Apple Silicon, Git and OpenSSL.

The local Docker stack includes the application, PostgreSQL metadata database, Redis, workers and Ollama. The Ollama service starts automatically; its AI model is downloaded separately. A paid provider can be used instead. No host Python or Node.js installation is required.

Windows quick start

Install Docker Desktop with Linux containers and PowerShell 7, then run:

git clone https://github.com/SlavchoSpasenoski/Bi-Dash.git
Set-Location Bi-Dash
pwsh -NoProfile -File ./scripts/start-local.ps1
docker compose -f compose.local.yaml exec ollama ollama pull qwen2.5-coder:7b

Open http://localhost:5001/setup to create your administrator. On later visits, open http://localhost:5001.

For Linux and macOS, clone the same repository and follow the guide’s shared initialization commands, which generate secrets and initialize a fresh database. The Windows script is also for fresh setup; use docker compose -f compose.local.yaml up -d for later starts.

The startup script generates local secrets in an ignored .env file. No AI API key is required for Ollama. Chat drafts SQL for manual use; Build dashboard creates and runs bounded read-only queries for PostgreSQL and MySQL/MariaDB. Review generated results before publishing.

This configuration is for local use at 127.0.0.1:5001; shared hosting requires separate security and network configuration. Windows/Linux-container operation has been verified; native Linux and macOS host installations still require platform validation.

An administrator can select a provider, enter its model ID and their own API key in Settings. Compatible services also accept a custom API base URL. Each installation uses its own provider account; no project owner’s key is bundled. Providers with other API protocols require an adapter.

See LOCAL_PROJECT.md for configuration, hardware limits, and validation, and BRANDING.md for the product identity.

Development checks

python -m unittest discover -s tests_local -v
docker compose -f compose.local.yaml config --quiet

The tests require the project’s Python dependencies, including cryptography. With the local Docker image built, run them without a host Python environment:

docker compose -f compose.local.yaml run --rm --no-deps --entrypoint python server -m unittest discover -s tests_local -v

Upstream and license

This project is based on an upstream source archive, with custom AI integration, local deployment, and branding. See UPSTREAM_SOURCE.json, README.upstream.md, and LICENSE. Original license and copyright notices are preserved. Original Bi-Dash contributions: © 2026 Slavcho Spasenoski, under BSD-2-Clause. See NOTICE and COPYRIGHT.md.

Local credentials, database storage, temporary files, and editor workspaces are excluded from Git.