VanillaGraph -- v0.1.0
VanillaGraph · Open source · MIT · Local-first

A local knowledge graph
where AI agents do the filing.

VanillaDB builds tools for agent memory and databases. VanillaGraph is our first — drop in documents, agents extract concepts and propose wiki articles, you approve. Your knowledge graph compounds entirely on your machine.

  • Four-agent pipeline: ingest → analysis → proposal → fileback
  • Knowledge graph with typed wikilinks and hub detection
  • MCP server — use your vault as a tool in Claude Desktop
  • OpenAI, Anthropic, OpenRouter, or fully-local Ollama
  • Hybrid BM25 + semantic search across your wiki
  • Tauri desktop app — Windows, macOS, Linux
View on GitHub
VanillaGraph · Open source · MIT · Local-first

Knowledge graph where
agents do the filing.

VanillaDB builds tools for agent memory and databases. VanillaGraph is our first — drop in documents, agents extract concepts and propose wiki articles, you approve. Entirely local, entirely yours.

View on GitHub
Product OS

Four agents,
one pipeline.

Documents go in. Structured wiki articles come out. Each agent in the pipeline has a narrow, auditable job.

01 — Ingest
Reader Agent
Parses PDF, markdown, and plain text. Extracts raw topics and entities.
02 — Analysis
Context Agent
Reads your existing wiki and knowledge graph to find gaps and links.
03 — Proposal
Writer Agent
Drafts structured articles with typed wikilinks. You review before anything is saved.
04 — Fileback
Commit Agent
Writes approved articles to your vault, updates the graph, logs to git.
wiki-vault/Transformer Architecture.md
--- title: Transformer Architecture tags: [deep-learning, attention, nlp] links: - [[Self-Attention Mechanism]] - [[Multi-Head Attention]] - [[Positional Encoding]] source: attention-is-all-you-need.pdf created: 2025-01-14 --- The Transformer is an encoder-decoder architecture that relies entirely on [[Self-Attention Mechanism]] to draw global dependencies between input and output...
demo.mov

See it run.

Drop a PDF. The pipeline runs. Articles appear for your review.

vanilla — pipeline
$vanilla drop attention-is-all-you-need.pdf
[ingest] reading 12 pages, extracting entities...
[ingest] ✓ 47 topics, 12 entities found
[analysis] scanning wiki-vault/ (84 articles, 312 edges)
[analysis] ✓ 3 gaps, 11 candidate links identified
[proposal] drafting articles...
[proposal] ✓ 3 proposals ready
▸ Transformer Architecture · new
▸ Self-Attention Mechanism · new
▸ Multi-Head Attention · update (3 links added)
$vanilla approve all
[fileback] ✓ 3 articles written to wiki-vault/
[fileback] ✓ graph updated — 3 nodes, 11 edges added
[fileback] ✓ committed: "add: Transformer Architecture batch"
$
Use Cases

Built for people
who think in documents.

VanillaDB fits any workflow where knowledge accumulates faster than you can file it.

Research
Academic & technical research
Drop in papers as you read them. The pipeline extracts claims, methods, and citations — and links them to what you already know. Build a personal research knowledge base that grows with your work.
Developer Docs
Living technical documentation
Feed in RFCs, ADRs, and runbooks. Agents link them into a coherent wiki. New team members can ask the MCP server questions about your system without reading every document.
Personal PKM
Second brain, no manual filing
Import meeting notes, articles, and highlights. Let agents handle the structure. You get a vault that actually stays organized because the work of organizing it is automated.
AI Workflows · MCP
Tool-use in Claude Desktop
VanillaDB ships an MCP server. Connect your vault to Claude Desktop and ask questions against your own knowledge. Your wiki becomes a retrieval-augmented tool your agents can reach.
Community

Built in the open.
Shaped by users.

VanillaDB is MIT-licensed and developed publicly on GitHub. Come shape what it becomes.

MIT
License — use it anywhere
v0.1
Early — your feedback shapes it
0→∞
Local-first, no accounts

We discuss roadmap decisions in the open, ship changelogs with every release, and welcome pull requests. If you use VanillaDB to manage your knowledge, we want to hear about it.

github.com/sanalvre/VanillaDB
Source, issues, PRs
Discord server
Chat, support, announcements
Changelog
What shipped in each release
Documentation
Setup, config, MCP guide
Future Products

What's coming.

VanillaDB is early. Here's what we're building toward — roughly in order.

Git sync & conflict resolution
Push your vault to a remote repo. Merge changes from collaborators without clobbering anything.
Soon
Multiplayer vaults
Share a vault with a team. Each member runs their own agent pipeline; proposals are synced and reviewed together.
Planned
Graph visualizer
Interactive canvas view of your knowledge graph — zoom in on clusters, follow wikilinks, discover hub articles.
Planned
Podcast & video ingestion
Transcribe audio and video files, then run them through the same pipeline. Your vault learns from everything you consume.
Planned
VanillaDB Cloud
Optional hosted backend for teams who don't want to manage their own git remote. Local-first stays the default.
Someday