Capture — Zero Friction, Any Source

Capture anything from anywhere — desktop app, Telegram bot, API, browser. AI classifies it automatically. You never decide where something goes.

The capture-first philosophy

Friction kills the habit of capturing. If you have to pause and ask yourself "is this a note or a task? Which folder does it go in? What tags should I add?" — you will not capture it. You will tell yourself you will do it later, and later never comes. The information is lost.

Synap is built on a simple principle: capture first, organize never. Every surface in Synap is designed to reduce capture friction to near zero. You send raw input — a thought, a link, a file, a voice message — and the system handles everything else. Classification, property extraction, relationship detection, entity creation. All automatic, all powered by AI.

Capture surfaces

Synap meets you wherever you are. There is no single "inbox" you have to remember to check. Instead, multiple capture surfaces feed into the same unified pipeline.

• Desktop app — The quick capture widget is always one keyboard shortcut away. Type a thought, paste a link, drop a file. It appears in your pod instantly.
• Telegram bot — Send any message to @SynapOS_bot. Text, links, files, photos, voice notes — anything you can send in Telegram becomes a structured entity in your pod. This is the fastest capture surface for mobile.
• API + webhooks — Programmatic capture from any source. Push data into your pod from scripts, automations, IFTTT, Zapier, or custom integrations. Every entity creation goes through the same AI classification pipeline.
• Browser extension (planned) — Right-click any page, highlight any text, save any tab. Coming soon.

What happens after capture

Every capture — regardless of source — enters the AI classification pipeline. This is a single pass that performs multiple operations at once:

• Detect entity type — Is this a note? A task? A bookmark? A contact? The AI examines the content and selects the most appropriate profile.
• Extract properties — Title, status, due date, URL, author, phone number. Structured data is pulled from unstructured input automatically.
• Create relationships — If the capture mentions a person, project, or company that already exists in your pod, the AI links them together.
• Create the entity — The fully classified, property-rich, relationship-linked entity is saved to your pod. Ready to appear in views, search, and dashboards.

The capture entity type

capture is a special entity type for things that have not been classified yet. When the AI pipeline cannot confidently determine a type — or when classification is deferred for batch processing — the input is saved as a capture entity. AI then processes it asynchronously and "promotes" it to the right type: a note, a task, a bookmark, whatever fits best.

You can also manually reclassify any capture by changing its profile. The capture type ensures nothing is ever lost, even when classification is uncertain.

Voice capture

Send a voice note on Telegram and watch it become structured data. The pipeline transcribes the audio, analyzes the transcript, classifies it, extracts properties, and creates an entity. Say "remind me to call Sarah about the Q3 budget next Tuesday" and you get a task entity with a due date and a relationship to the person entity "Sarah" — all from a 5-second voice message.

Batch capture

Synap handles bulk input just as well as individual captures. Import a CSV of contacts, paste multiple URLs at once, or forward an entire email thread. Each item goes through the same classification pipeline independently. A CSV with 200 rows becomes 200 fully typed, property-rich entities in minutes.

Related guides