Help center

Short, honest answers about what yonja.ai does today — and how to get unstuck.

Getting started

Sign up, verify your email, and walk through onboarding: your brand, sector, goal, voice tone, links and competitors. yonja.ai then runs a strategy analysis (it takes a couple of minutes) and builds your strategy file — the document every content draft is grounded in. You can re-run the analysis from the dashboard whenever your positioning changes.

The Stage, Windowed and Desktop modes

After login you land on the Stage: a command bar, a live brief of what needs your attention, and a constellation of your modules with their real status. The switcher in the top bar changes how the same product is presented — Windowed puts each module in a single window, Desktop gives you draggable windows, icons and a taskbar. Your choice is remembered across devices.

Talking to Yonja

Yonja is the assistant behind the command bar and the chat threads. With Google connected it can summarize your inbox, list and create calendar events, draft emails (saved to your Gmail drafts, never sent by itself) and read Drive documents. It can also tell you what's waiting in your approval queue and produce content drafts from your strategy file ('draft a post about our launch') — drafts always land in the queue, nothing publishes without your approval. Ask 'what can you do?' for the current list. It refuses actions it cannot actually do — if it claims something happened, that action really happened and has a verifiable ID.

Connecting Google (Gmail, Calendar, Drive)

Settings → Connect Google walks you through OAuth consent. Identity access (mail, calendar, files) is a separate connection from publishing access (YouTube channel) by design — Brand Accounts need the separate flow. You can disconnect either at any time from Settings; disconnecting removes the stored tokens.

Connecting X and YouTube

Settings lets you connect an X account and a YouTube channel for publishing. Nothing is ever posted without an approval: drafts wait in the content queue until you approve them, and published items store the real platform post ID so you can verify them on the platform itself.

The content pipeline

Generate drafts from your strategy file, review them in the queue (amber means a decision is waiting), approve or reject, then schedule. A daily scheduler publishes approved items — Mode A emails you the ready-to-post content, Mode B posts directly to the platform (YouTube uploads today). Failed publishes show the real error and can be rescheduled; you also get an email alert when a scheduled publish fails.

Avatars and video generation

The Avatars module generates images and short clips through Higgsfield, using your avatar photo as a visual reference (exact identity match is limited by the provider API today). Video generation costs credits from a monthly free-tier allowance; if the pool runs out you get an honest error, never a silent failure. A brief can also become a 2-3 scene storyboard: each scene is generated as its own clip and joined into one video, and if generation is interrupted, resuming never re-buys a scene you already paid for. Generated media lives in your workspace storage and shows up in the approval queue with an inline preview before anything is published.

Plans, limits and fair use

The free tier includes monthly analysis runs, daily chat messages, monthly video credits and daily posts. There are also burst limits per minute to keep the service stable for everyone. When you hit a limit the response tells you which one and when it resets. Power users can add their own API key (BYOK) in Settings — your key is encrypted at rest and used only for your requests.

Your data: export and deletion

Settings → Your data lets you download everything yonja.ai holds about you as a single JSON file (connection secrets are never included). Account deletion is two-step — you type your email to confirm — and permanently removes your workspaces, content, connections, generated media and the account itself. Anonymous usage accounting (no personal data) is retained.

Troubleshooting and reporting problems

If an analysis or chat fails with a capacity message, wait a few minutes — free-tier AI providers rate-limit under load, and the system falls back automatically. If something looks genuinely broken, use the report form below. API responses carry an x-request-id header; if you can see it (browser dev tools → Network), include it in your report so we can find the exact request in our logs.

Report a problem

Something broke or looks wrong? Tell us. If you have an error message, paste it in — it helps us find your request in the logs.

Back to home

yonja.ai