Adding Personal Credits
To add your own credits for personal use:- Open Your Profile › Billing and click Add Credits
- Scroll to the bottom of the page
- Select your desired credit amount and complete payment
How credits are consumed
Credits are deducted in real time as you interact with your Digital Twin. The core unit is the token — a small piece of text, roughly four characters. Everything the model reads and writes on a turn is counted in tokens, and the total is converted into credits.Standard conversations — one flat rate
A standard dialogue turn (one conversation model answering you directly) is billed at a flat rate: 10,000 tokens = 1 credit, counting input and output together, with a 1-credit minimum per turn and a single round-up per turn. There is no separate input rate, output rate, or per-model surcharge for standard turns — one simple rule covers them all.With a few very-large-context models, once a request gets very long the whole request is charged at a higher rate — another reason to keep prompts lean.
Agent Mode — true cost, itemized
When a dialogue runs in Agent Mode, one answer involves several AIs — the conversation model, the manager coordinating the work, and each helper it brings in. The flat 10,000-tokens-per-credit rule does not apply to these turns. Instead you’re charged what the work actually cost, plus a small platform margin. Here is the whole billing model, in order:- Every model that worked on the turn is counted, in three groups: the conversation model that wrote your answer, the manager that decided what to do, and each helper it dispatched.
- Each one is priced at its own listed rate. A cheap, fast helper costs what a cheap, fast helper costs — it is not billed at the premium conversation model’s rate. Text the provider served from its cache is priced at the cheaper cached rate.
- A small platform margin is applied on top of that cost. It is the same margin on every model, and it is shown to you in the itemized bill.
- Everything is added together and rounded up once, at the end of the whole turn — never per helper. Ten helpers do not mean ten roundings.
Why Agent Mode is priced this way. An agent turn routinely reads hundreds of thousands of tokens of cached context across a dozen model calls. The flat rate used for standard turns would charge you far more than the work cost on those turns — so Agent Mode drops the flat rate and bills the real cost instead. You pay for the team you actually used.
Failures and cancellations: a turn that errors before any real work is never billed, and a turn you cancel is never billed. An Agent Mode run that fails partway after doing real work bills only the work actually done — exactly what Usage Details itemizes.
Non-text actions — metered separately
Talking to an on-screen avatar
When your Digital Twin answers as a lifelike animated avatar (the Anam avatar) in a one-to-one Convo Mode session, your credits are metered the same way as any other spoken session: on the tokens of the conversation, totalled across the session and charged when the session ends. The avatar does not add a per-minute line item to your balance in Convo Mode — you won’t find “avatar minutes” on your bill. So an avatar exchange and a voice-only exchange of the same length cost you about the same. What makes the avatar the more expensive habit is how it gets used:- Spoken turns cost more than typed ones. Audio tokens are priced well above text tokens by every voice provider, so any voice conversation — avatar or not — costs more than the same conversation typed.
- Avatar sessions run long. People stay in a conversation with a face on screen far longer than they stay in a chat box, and the whole session is one bill.
Sending your twin into a meeting
Billing works differently when your Digital Twin joins a Zoom, Google Meet, or Teams call as the Meeting Agent, rather than talking with you one-to-one in Convo Mode. For every minute the meeting bot is in the call, credits are metered per minute for the avatar, the meeting bot itself, and live transcription of the call — on top of the normal per-turn rule for whatever the twin actually says when it speaks. Those per-minute charges are what the avatar, meeting-bot and transcription providers charge us, times one and a half, rounded up to a whole credit once at the end of the meeting — not every minute. A scheduled meeting is billed from the moment the bot joins the call, not from when it was scheduled, and ending the meeting — including by asking Pria to leave — is billed the same as any other ending. See Meeting Agent for the full picture.If your organisation supplies its own Anam credentials, Anam bills that account directly for avatar minutes — in both Convo Mode and meetings — which is a separate bill from your Pria credits. Digital Twins on Praxis-managed Anam access have it all consolidated into credits. Either way, the meeting bot and transcription are always Praxis-managed, so their per-minute charges bill in credits regardless of whose Anam account the avatar uses. See Anam Avatar.
At-a-glance cost table
The rough relative costs (lowest → highest) for common actions:Star ratings are relative — they convey order of magnitude, not exact rates. Your balance and usage history live under Your Profile → Billing; see Plans & Credits for the pricing rules.
Where credits live — your organisation, each Digital Twin, and you
Pria’s credits sit at three levels:
Which pool pays is set by the twin’s Credit Management mode, not by which balance happens to be positive:
- Pool Credits — conversations on that twin draw from the twin’s pool. Depending on the account’s transfer policy, the pool can be refilled from the account reserve automatically when it runs low, only by an admin, or never.
- Personal Credits — each user pays from their own balance on that twin. Users on an unlimited monthly plan converse without their balance being deducted.
Low-balance behaviour
Pria tries to warn the right people before credits run out so you’re never mid-conversation when the lights go off.Alerts
- Admin notifications — when a twin’s shared pool drops below its threshold, a courtesy email goes out so the people who can top it up know, even if they aren’t signed in. (It’s skipped when the parent account is set to refill the pool automatically.)
- Credits display — your balance (or the twin’s pool balance, for admins of pooled twins) shows next to your current Digital Twin in the Digital Twin tray: select the arrow next to your current Digital Twin in the top bar. When the balance reaches zero, the arrow shows it in red.
Automatic refill (pooled twins)
If the parent account’s transfer policy is set to auto-refill on low credits, Pria automatically moves a batch of credits from the account reserve into the twin’s pool when it runs low — conversations continue without interruption for as long as the account reserve holds out. Admins configure the transfer policy under Credit Management.When credits hit zero
If the pool that pays for your conversations runs out:- In-flight response — the running response finishes (Pria doesn’t cut you off mid-sentence), but the next prompt is blocked.
- New prompts show a clear out-of-credits message with a direct path to Add Credits (personal-credit twins) or a note that your Digital Twin has no credits left (pooled twins — contact your admin).
- File uploads, image generation, voice are blocked the same way rather than producing partial output.
- Read-only features (browsing past conversations, viewing files) continue to work — you only lose the ability to spend new credits.
Optimization tips
A small handful of habits can cut credit consumption dramatically. See Credits Optimization for the full guide; the short version:- Lower Remember History when long context isn’t needed (quick lookups don’t need 10 turns of history).
- Tune your retrieval mode — Disabled or RAG is cheaper than RAG + KAG Fusion; pick the right mode for the question.
- Skip Indexing on URL imports you only want to download, not embed.
- Prefer text over Convo Mode for fact lookups — voice is convenient but more expensive per turn.
- Use Skip Indexing for files you only need to store and open yourself (not search) — they skip the knowledge-processing pipeline entirely, so they cost nothing to index.
- Sanitize noisy segments so they’re shorter to embed and retrieve.
- Avoid HD image gen unless you need print-grade output.
- Personal accounts: leave Tools Guidance off unless you need Pria’s detailed tool instructions — see Tools Guidance.
Related
- Plans & Credits — Pricing rules, caching discounts, Single-Shot and Agent Mode billing.
- Credits Optimization — Practical tips to reduce credit consumption.
- Credit Management (admin) — Account / Digital Twin allocation and transfer policies.
- Billing (admin) — Payment methods, invoices, billing contacts.
- Edit Profile — Where the Add Credits and Billing tabs live.
- Sidebar & History — Remember History, the main lever for conversation cost.
- AI Models — Provider price tiers and reasoning support.