Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Interactive 3D visualization of knowledge graphs generated by Microsoft GraphRAG. Explore entities, relationships, and communities with immersive visual analytics.
| Date | Stars |
|---|---|
| 2026-07-31 | 711 |
| 2026-08-02 | 712 |
| 2026-08-06 | 712 |
| 2026-08-12 | 711 |
| 2026-08-18 | 712 |
| 2026-08-19 | 712 |
| 2026-08-21 | 713 |
| 2026-08-28 | 714 |
| 2026-09-01 | 717 |
| 2026-09-02 | 718 |
| 2026-09-04 | 719 |
| 2026-09-08 | 720 |
| 2026-09-11 | 721 |
| 2026-09-18 | 722 |
| 2026-09-20 | 722 |
Today
— stars today
This week
+1 stars this week
This month
+9 stars this month
Momentum
0.0
growth rate 0.14%/day
<p align="center"> <img src="https://raw.githubusercontent.com/lyon-industries/.github/main/profile/lyon-industries-banner.svg" alt="Lyon Industries" width="100%"> </p> # GraphRAG Workbench A local workbench for turning documents into an inspectable 3D knowledge graph. GraphRAG Workbench wraps [Microsoft GraphRAG](https://github.com/microsoft/graphrag) with local document preparation, indexing controls, a live terminal, project management, and the original Three.js graph. It is a dedicated open-source desktop-style web app: clone it, run it, and configure local or cloud models inside the Builder. https://github.com/user-attachments/assets/1f588a45-07ca-4953-92ed-fc888fe28cff ## Version 2.0  *Inspect entity evidence, community hierarchy, and strongest relationships without leaving the live graph.* - Microsoft GraphRAG 3.1.0, pinned with `uv` - per-build Local / Ollama and Cloud / OpenAI presets configured in the interface - full-screen 3D graph with search and community isolation - contextual Inspector for selected entities and their strongest connections - local Projects sheet for naming, loading, renaming, deleting, files, statistics, builds, and terminal output - native GraphRAG output-directory import plus portable Workbench bundle import/export - cancellable server-owned indexing with persisted workflow status; builds survive closing the Builder - live constellation population while a build runs: entities and relationships appear as extraction completes, communities as clustering completes - engine-log surfacing with fast failure on fatal provider errors (quota, authentication, missing model) - text-backed PDF validation and transactional file removal - no account, hosted database, or remote document service  *Manage projects and monitor indexing while retaining the graph and selected-entity context.* Chat is intentionally absent from 2.0 while its next interaction model is designed. ## Requirements - Node.js 20.9 or later - pnpm - Python 3.12 - [uv](https://docs.astral.sh/uv/) - no model provider is required before first launch; the Builder walks through Ollama or OpenAI setup ## Install ```bash git clone https://github.com/lyon-industries/graphrag-workbench.git cd graphrag-workbench pnpm install uv sync --frozen pnpm dev ``` Open [http://127.0.0.1:3000](http://127.0.0.1:3000). Development and production commands bind to the local interface. ## Configure builds in the interface Open **Projects → Build providers**. Both presets can remain configured at the same time: - **Local / Ollama** detects the Ollama installation and service, links to the official installer when absent, and pulls the selected completion and embedding models with visible progress. The default preset is Gemma 4 plus EmbeddingGemma at 768 dimensions. - **Cloud / OpenAI** accepts the API key and model names in a modal. The key is saved in the workbench-wide `.graphrag/providers.json`, restricted to the operating-system account (`0600`). It is outside every project, excluded from Git, and never returned to browser code. The Ollama and OpenAI bindings are workbench-wide settings, reused by every GraphRAG project. Each indexing run has an explicit **Run with Ollama** or **Run with OpenAI** command. Ollama avoids provider token charges and keeps model processing on the workstation; OpenAI is normally faster and can use a stronger extraction model. OpenAI indexing sends document content to the configured provider and can consume substantial model tokens. During a build the engine log is tailed into the Terminal. Fatal provider failures such as exhausted quota, a rejected key, or a missing model
Excerpt of 6,688 characters
Read on GitHub
14
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:b4d1e8cf692df2f9, topic:graphrag