SuperAPI AI Directory

Anarlog

Open-source, local-first AI meeting notepad (formerly Hyprnote) — records and transcribes meetings on-device and turns raw notes into structured summaries using local models.

Open website

About this tool

Anarlog is an open-source, local-first AI meeting notepad positioned as an alternative to Granola, formerly known as Hyprnote and later renamed Char. Its core goal is to let users stay focused during meetings while keeping meeting records always under their own control.

Purpose Anarlog is aimed at individuals and teams who need to record private meeting content. It emphasizes privacy and data autonomy: meeting notes, transcripts, attachments, and recordings are stored on the user's device by default, with no mandatory cloud uploads and no locking of data inside the application.

Main capabilities

  • Bot-free participation: Anarlog does not join meeting calls as a bot, stays hidden when screens are shared, and works silently from the desktop.
  • Local-first: Notes, transcripts, attachments, and recordings stay on the device by default.
  • Editable AI summaries: Transform raw meeting notes into structured, editable summaries.
  • Autonomous AI stack: Run models on-device, bring your own API key, or use the officially provided AI service.
  • Dictation: Built-in dictation turns spoken ideas into text and can output it in the apps the user already uses. The official example shows the effect of polishing colloquial content (including filler words such as "um," "uh," and "like") into fluent text.
  • Open-source transparency: Uses a GPL codebase, so anyone can inspect how meeting memory is handled.

How to use Anarlog runs on the desktop, letting users record meetings without having a bot join the call. Users can choose to run models on-device, use their own API key, or use the official AI. Dictation can be used directly within Anarlog and output text to other apps. Officially, three paths are offered: a free local start, Pro cloud features and collaboration, and Enterprise organization-level controls.

Use cases Suitable for privacy-conscious meeting recording scenarios, such as team meetings involving sensitive conversations, work environments that require offline or local recording, and users who want to control their own AI models and keys. The official list includes enterprise users such as Databricks, Cloudflare, Amazon, Meta, Y Combinator, Palantir, Apple, Disney, Adobe, Wayfair, Bain & Company, and McKinsey & Company.

Limitations and information not yet provided Official text does not provide specific pricing, a complete list of feature differences between versions, supported operating systems, model names and performance metrics, transcription language supportor scope details; this information is not yet available.