Synaplan now plans MCP tool calls right into its DAG routing: connect your own MCP servers and the AI gathers data from your systems upfront, before it answers. And every Synaplan – cloud or self-hosted – ships with a free, built-in MCP server.
AI with Nextcloud, done sovereignly: the new Synaplan AI Assistant app (v1.4.0) brings chat, your own knowledge (RAG), memories and /pic + /vid image and video generation right into Nextcloud — and saves results straight back to your files.
Ask Synaplan one complex thing and it doesn't fire a single AI call — it sketches a small plan, a directed acyclic graph of steps, and runs them right in front of you. Here's why that's a gift if you self-host, shown with a simple calendar-invite example.
Summarise threads, translate, draft replies, classify mail and save to your knowledge base — without leaving Outlook. And you can point Synamail at your own self-hosted Synaplan server.
AI agents write a large share of our code — which is exactly why every commit runs a no-shortcuts CI pipeline: PSR-12 linting, PHPStan, vue-tsc, 549 Vitest tests, Playwright across two browsers, OIDC login flows, and a full Docker image build. Nothing ships to production until every check is green. A look behind the scenes of our open-source AI platform.
Today’s best language models live in US or Asian data centres — but your documents do not have to follow them there. With a sovereign, locally managed vector database, your team can keep every memory and every document on its own infrastructure while still using GPT-5, Claude, Gemini, or Qwen as the inference engine. Here is why that split matters, especially for larger corporations.
Big Steps Added Fable 5 back and welcomed Sol, Terra and Luna. Displaying a token taxameter showing your spendings. Int…
What's Changed * Refactor: Replace GuestFeatureGateModal with GuestHintPopover by @orgaralf in https://github.com/metad…
We have added a cool MCP call during prompting for your pleasure: https://www.youtube.com/watch?v=mAqRNKK9HBw Some bug…
Join our Discord
Chat with the team and community — questions, feedback, and feature discussions.
