Getting started

Quick AI Chat (available when no file selected)

Image Creation

Audio Creation

File Text Extract (when a Drive file is selected)

Permissions & privacy

• Per-file permission model. Flowmata AI requests access only when you initiate an operation that requires a file. Granting permission is a Drive dialog; the app does not scan or touch files automatically.
• No persistent file crawl. Flowmata AI has no access to your entire Drive.
• Data handling and retention. Generated images and audio are stored on Flowmata AI servers for a limited time (48 hours) and then removed. Metadata necessary for history and billing is stored securely. The exact retention policy and data handling details are available on the Flowmata AI website.
• Security. Traffic between your browser and Flowmata AI is encrypted. Authentication is handled via Google OAuth and per-file Drive permissions.

Credits, pricing & promo

• Credits visible in the sidebar. Top of every Flowmata AI tools shows your current balance and quick-buy buttons for 10, 20, and 50 credits at $10, $20, and $50 respectively (with discounts during promotion).
• Promo credits. New users receive 5 promo credits on first run. Promo credits are consumed first. When depleted, paid credits are used.
• Per-model pricing. Each model consumes credits at different rates. At launch, Flux.1-schnell (image), MeloTTS (TTS), and Llama 3.1 8b (chat) are available as free. During the promotional period, image generation prices are extremely low (example: 0.01 credits per generation). Model-specific pricing and charging rules are documented on the Flowmata AI site and in the sidebar model info.
• Billing & receipts. Purchases are charged via Dodo Payment service.

Files and where they are saved

• Default save location. Generated files are saved into the Drive folder that was open when the user launched an AI task. But in certain cases file will be saved in the folder “Flowmata AI” that is created upon add-on first run (this happens when Flowmata AI doesn’t have access to the file that was last selected, or the first file in the folder; You need to give permission to access this file first).
• Save notification. After a file is saved, the app displays a confirmation message with the folder and file name.
• Editable filename and file type. After generation, Flowmata AI sets a generic filename (e.g., Image - ai_img_[date]_[id].jpg or Audio - TTS - tts_audio_[date]_id.mp3) which you can rename like any other Drive file.

Job status and long-running operations

• Image and audio generation can sometimes take longer than a few seconds depending on model load and complexity.
• Polling flow. If a generation takes more than 30 seconds, use button Check Status & Get Results to query job status. The sidebar provides both direct-result display and a history where you can re-download generated files.
• History retention. Completed generation results are downloadable for 48 hours; after that, the files are automatically purged from Flowmata servers but remain in Drive if saved there.

Models & engines (at launch)

• Image models: Flux.1-schnell (fast, free), Stability.ai Stable-diffusion-xl-lightning (balanced), Leonardo Lucid-origin (artistic), Leonardo Phoenix-1.0 (photo-real options).
• Audio models: MeloTTS (ultra-simple, free), Deepgram-1 (higher quality, multiple voices).
• Chat models: GPT OSS 120b, GPT OSS 20b, Llama 4 scout 17b, Llama 3.3 70b, Llama 3.1 8b (free), Mistral small 3.1 24b, Deepseek r1 distill qwen 32b, QwQ 32b, Gemma-3-12b.

Note: Model availability and pricing may change as Flowmata integrates additional providers and optimizes costs. Model-specific behavior (image aspect ratios, sampling options, TTS voice lists, etc) are visible when model panels are opened in the sidebar.