Plans & AI options
Agent28 charges for one thing: managed AI, measured in AI tokens. Everything that runs on your machine is unlimited on every plan, and two of the three ways to run AI cost nothing at all.
The tiers
| Plan | Price | Managed AI tokens / month | Summary |
|---|---|---|---|
| Free | $0 | None | Bring your own API keys or use ChatGPT / Claude. Unlimited renders. |
| Plus | $29 / mo or $299 / yr | 5,000 | Managed AI with no keys, the full caption and motion-graphics catalog, meme, music and SFX search, AI voice narration. |
| Pro | $59 / mo or $599 / yr | 15,000 | Everything in Plus, the best models on every task, style cloning, video recreate, licensed web-footage suggestions. |
| Studio | From $99 / mo | 30,000 to 300,000 | Everything in Pro, custom narration voices, priority support. Pick a monthly credit pack; above $99, 300 tokens per dollar. |
Plans are managed under Settings → Billing, which opens the Stripe billing portal for receipts, payment method, plan changes, and cancellation. Yearly billing is ten months for the price of twelve. Teams can contact sales for volume pricing and seats.
How AI tokens work
- Tokens are spent only when the in-app chat runs on Agent28's managed models. The cost of a turn shows in the usage popover under that turn.
- The plan allowance is spent first, then any bonus tokens. Allowances reset each billing period; bonus tokens do not renew and are only spendable while a paid plan is active.
- When the allowance runs out, managed AI pauses with the message “AI token limit reached”. If you have your own API keys saved, the chat keeps working on those. Otherwise upgrade, or wait for the reset.
- Never metered: your own API keys, ChatGPT, Claude, Claude Code, rendering and export, local transcription, person matte, and scene indexing.
Your own API keys
Under Settings → API Keys you can save keys for OpenAI, Gemini, Claude, Kimi, and ElevenLabs. Keys are stored in your system keychain and sent only to that provider; the page tells you whether keychain storage is available on your system. Each key is verified when you click Save.
- On Free, a key is what makes the in-app chat work. Usage is billed by the provider and is unlimited as far as Agent28 is concerned.
- On a paid plan, keys take over when the allowance is exhausted or managed AI is paused.
- With your own key, the LLM-driven Pro features (style cloning, video recreate, best-model access) are not tier-gated. Services Agent28 pays for itself stay gated: meme and GIF search, music and SFX search, licensed web footage, voice narration, and the premium catalogs.
Features by tier
| Included from | Feature |
|---|---|
| Plus | Voice narration |
| Plus | Meme & GIF search (KLIPY) |
| Plus | Music & SFX search (Freesound) |
| Plus | The full caption & motion-graphics catalog |
| Pro | Best-model access on every task |
| Pro | Style cloning (Style Studio) |
| Pro | Video recreate |
| Pro | Licensed web-footage suggestions |
| Studio | Custom narration voices |
| Studio | Early access to new styles & templates |
| Studio | Priority support |
A gated feature explains itself when you try it: “Meme & GIF search is included with Plus and above. Upgrade your plan.”
Choosing models
Settings → Models maps every AI feature to one of three tiers, each with a model picker that defaults to Auto (recommended). Auto picks the best model among the providers you have keys for, or the managed model on a paid plan.
| Tier | Used for |
|---|---|
| Complex | Deep reasoning: cut planning, motion-graphics direction and code. |
| High | Drafting and editing: chat, story planning, critique, editorial edits. |
| Low | Fast utility: research, captions, ingest, validation. |
- Generation defaults: separate pickers for Image generation (backdrops, cutout subjects) and Video generation (generated b-roll).
- Advanced — per-task overrides: move a single feature (for example “Caption agent” or “Motion graphics code”) to another tier, or pin it to an exact model. A pinned model wins over everything.
- When a provider has no key, its models are greyed out. The line under each picker shows what Auto currently resolves to.
AI Credits, Billing, Free Credits
- AI Credits: tokens left this billing period, used, plan allowance, bonus tokens, and the reset date, with a Refresh button. On Free it explains the two free options instead.
- Billing: current plan, status, renewal date, and the Manage billing button for the Stripe portal.
- Free Credits: follow Agent28 on Discord, X, or Instagram and each adds 50 bonus tokens, once per platform. Bonus tokens sit after the plan allowance and are spendable only on a paid plan.
Tip: Current prices and allowances are also on the Pricing page.