Task Credits

1. What credits are

Credits are the unit that meters AI capacity used on PMTSoul, the reward for contributing to the platform, and the currency of the Task Market. Every change is bound to its source and is auditable end to end.

Credits play three roles:

• A meter — they measure the AI resources a task run consumes (model calls, tool calls, knowledge retrieval), deducted from your company or personal credit account.
• A reward — they recognise contributions to the platform (creating tasks or skills, contributing knowledge, confirming outcomes) and are added to your reward account.
• A currency — they price business capabilities in the Task Market and settle creator earnings.

Core rule: spending credits and reward credits are separate accounts, and transfer is one-way. This exists to prevent self-dealing — running your own tasks to consume credits while earning them back as contributions.

2. How you get credits

• Subscription — a monthly allowance of base credits, plus a bonus on your first subscription. When the allowance runs out you can top up; top-ups land immediately. Exact amounts are on the pricing page.
• Contributing — creating tasks or skills that other companies use, contributing knowledge, or submitting feedback earns reward credits.
• Task Market earnings — revenue share for creators who publish tasks or skills (in cash or credits).

📌 Amounts and prices live on the pricing page; this page deliberately does not repeat them. The pricing page is driven directly by the backend price authority, so a price change takes effect immediately. Copying the numbers here would let the two drift apart — silently.

3. How credits are spent

• Task runs — a digital employee spends credits when it runs a task.
• Redemptions — extra compute capacity, premium skills, priority support.
• Task Market purchases — installing a paid task or skill.

4. How a task run is metered

You install tasks from the Task Market. Each run — whether in full-agency mode or run manually — spends that task's credits.

What is metered: one task run = model calls + tool calls + knowledge retrieval. Today the platform meters per task at a fixed credit cost; the Task Market shows each task's cost, and it is deducted at run time.

• Every task is metered separately, so you can see which task costs the most.
• Failed runs are not charged — aborts, timeouts and system errors.

5. Contribution rewards

Reward credits recognise contributions to the platform. They go to your reward account, can be redeemed for benefits, and cannot be converted back into spending credits.

• A task you created is installed by another company — high reward, per install.
• A skill you created is invoked — high reward, per invocation.
• You confirm an AI judgement — medium reward per confirmation, subject to a cap.
• Knowledge you uploaded is actually used by a task — medium reward, per citation.
• You report an execution issue or edge case — low reward, once reviewed.
• You review a Task Market listing — low reward, subject to anti-gaming checks.

6. Redeeming credits

• Compute capacity — extra model capacity for peak demand.
• Task Market purchases — credits offset the install cost of a task or skill.
• Premium capabilities — premium skills, priority support, bespoke training and custom digital employees.

7. Validity and rollover

• Base credits — issued each subscription period. Unused credits roll over automatically for up to one year; they are cleared if the subscription lapses continuously.
• Bonus credits — valid for one month from the day they land, then expire.
• Reward credits — roll on a calendar-year basis, with a reminder 30 days before annual expiry.
• Top-up credits — carry their own validity from the date of purchase; expired credits are not refundable.

8. Statements and audit

You can review your credit ledger under Me — System — Statistics: the type of each change (spend / refund / net), the amount, the task it is bound to, when it happened, and the linked execution record.

Administrators can see organisation-level reports: credit spend by department, average cost per task, and resource efficiency per digital employee.

Every credit change is written to a tamper-evident audit log, connected to the product's audit chain — who initiated it, on what basis, which tools were called, what a human confirmed, the result, and the compliance verdict.

9. Changes to these rules

PMTSoul sets and interprets these rules. Prices and credit rules may change as the service evolves; we will notify you 30 days before any change takes effect.