Plans & Pricing
Compare plans and find the right token allocation for your needs.
Token System
Tokens are the currency that powers every AI interaction in Primio. Understanding how they work helps you build more efficiently and avoid surprises.
How tokens work
Section titled “How tokens work”Every interaction with the AI consumes tokens. This includes sending a message, the AI reading your project files, generating code, and fixing errors. The total cost of an interaction depends on how much the AI reads and writes.
Different models consume tokens at different rates:
| Model | Token cost | Best for |
|---|---|---|
| Sonnet | Lower per interaction | Balanced work (features, iteration, simple tasks) |
| Opus | Highest per interaction | Complex tasks, best code and UI quality |
| Auto | Varies | System picks the optimal model for each request |
See Model Selection for detailed guidance on when to use each model.
Publishing costs
Section titled “Publishing costs”Builds and exports have fixed token costs, independent of which AI model you use:
| Action | Token cost |
|---|---|
| Web hosting (PWA) | Free |
| PWA export | 20,000 |
| Android APK | 20,000 |
| Android AAB | 40,000 |
| iOS IPA | 40,000 |
See Publishing Overview for details on each build type.
Minimum for new projects
Section titled “Minimum for new projects”You need at least 20,000 tokens to create a new project. If your balance is below this threshold, the app will prompt you to upgrade your plan before you can start a new project.
Token rollover
Section titled “Token rollover”Low token warning
Section titled “Low token warning”When your balance drops below 100,000 tokens, a warning appears in the chat input area. Click Add tokens to open the pricing modal directly.
The warning can be dismissed for the current session. It reappears when you reload the page.
Your token balance refreshes automatically and updates immediately when a task completes.
Understanding token consumption
Section titled “Understanding token consumption”This is the most common source of questions. Here’s what drives token usage and how to stay in control.
Why tokens can spike unexpectedly
Section titled “Why tokens can spike unexpectedly”AI error loops. When the AI encounters a compilation error, it tries to fix it automatically. Sometimes it enters a loop of repeated fix attempts. If you see the AI making similar changes over and over, interrupt it by sending a new message with clearer direction, or use Rollback to revert to a known good state.
Complex refactors. Large operations like splitting files, restructuring navigation, or reworking state management consume more tokens because the AI reads and rewrites many files in a single pass.
Vague prompts. Unspecific requests like “fix everything” or “make it better” cause the AI to explore broadly, reading many files and making scattered changes. This burns tokens without focused results.
Long conversations. As your project grows, each interaction includes more context about your codebase. More context means more tokens per message.
How to minimize token waste
Section titled “How to minimize token waste”Be specific. Clear, focused prompts consume fewer tokens. Instead of “improve the design,” say “change the header background to dark blue and increase the font size to 18px.” See Prompting Fundamentals.
One feature per message. Don’t batch multiple unrelated requests into a single prompt. Each focused request is cheaper and produces better results.
Use Sonnet for simple changes. Text updates, color changes, and small tweaks cost less on Sonnet than on Opus. Switch models as needed. See Model Selection.
Rollback early. If the AI is heading in the wrong direction, rollback immediately rather than sending multiple corrective messages. Each correction costs tokens.
Use @file references. Point the AI at specific files instead of letting it scan the whole project. This reduces the context it needs to process. See Attaching Context.
What to do if tokens are depleted unexpectedly
Section titled “What to do if tokens are depleted unexpectedly”- Check your token balance in the chat input area or in Personal Settings under Plans.
- Review your recent conversation to identify what consumed the most tokens (look for long AI responses, error fix loops, or broad refactors).
- Upgrade your plan for more tokens via the pricing modal.
- If you believe there’s an error, contact support at support@primio.dev with your project details.
Monitoring your balance
Section titled “Monitoring your balance”Your token balance is visible in two places:
- Chat input area — always visible while you work
- Personal Settings — under the Plans section, with your billing period dates
The balance updates automatically.
Next steps
Section titled “Next steps”Prompting fundamentals
Write efficient prompts that get better results with fewer tokens.
Getting the Best Results
Practical strategies for saving tokens and working effectively with the AI agent.