What is Kompas AI?
Kompas is a local-first desktop app that captures notes, documents, highlights, web pages, and voice notes into a personal knowledge graph.
It organizes work into sessions that retrieve related notes, decisions, and passages to support progressive thinking and connected writing.
Users can bring their own LLM API keys (OpenAI, Anthropic, Claude) or run local models via Ollama; realtime transcription requires an OpenAI key.
Private hosted sync and hosted models are available for cross-device access and managed AI usage, while the desktop mode keeps data local and supports JSON export.
Kompas preserves data ownership and does not use user content to train models; the desktop app is open source under the MIT license.
Target users include essayists, researchers, founders, and Obsidian/Logseq users who need retrieval-driven drafting, organized knowledge graphs, and exportable archives.
Kompas AI user reviews
Would you recommend Kompas AI?
Kompas AI's key features
-
Local-first desktop app with offline editing and local data storage
-
Supports bring-your-own LLMs: OpenAI, Anthropic, Claude/Codex, and local models via Ollama
-
Pro cloud with cross-device sync and hosted private graph storage
-
Built-in managed AI (Pro Full) with included usage limits and optional top-ups (no personal API key required)
-
Capture-to-session workflow that ingests documents, highlights, web pages, and voice notes into a connected session/graph with retrieval into drafts and JSON export
Kompas AI use cases
-
Draft research-backed articles and long-form essays using Kompas's retrieval-driven drafting: capture notes, highlights, web pages and voice into a local-first knowledge graph, run connected writing sessions that surface related material and refine drafts with offline LLMs while keeping your data private
-
Organize and maintain project documentation and meeting notes by grouping resources into sessions, automatically retrieving context from your personal graph, privately syncing across devices with self-hosted sync and exporting a versioned knowledge archive for handoff
-
Build a secure, searchable research library for academics or students by clipping web pages and highlights into Kompas, using offline LLM note-taking to summarize sources, generate connected outlines and progressive drafts, and export the privacy-first knowledge archive for sharing or backup
Who is it for?
-
Creative writers
-
Data researchers
-
Knowledge workers
-
Academic students
-
Productivity enthusiasts