Who it is forA team that cannot send documents to a hosted model.
The problemHosted LLMs mean documents leave your environment.
The outcomeWe Knora answering from your documents using a local Ollama model.
What you need first
- Docker with Docker Compose
- An LLM provider key, or a local Ollama model
- Optional Neo4j, MinIO or Langfuse via compose profiles
Steps
- Run
ollama serve > /dev/null 2>&1 &first - Set the Ollama embedding model name and OLLAMA_BASE_URL as the Configuration docs describe
- Start the stack with
docker compose up -d - Pick the Ollama model in the model settings
Prompt for your agent
Configure WeKnora to use my local Ollama for chat and embeddings (OLLAMA_BASE_URL=<url>), restart with `docker compose up -d`, and confirm a test question works offline.
Caveat. See the Configuration docs for RAM notes on local models.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for Tencent/WeKnora and "Keep everything local with Ollama models"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer