跳转到内容
Wetel Docs

Pricing & Plans

此内容尚不支持你的语言。

Wetel’s commercial model is still evolving, so this page separates what is real and enforced today from what is not built yet. See Getting Started to start building in the meantime.

How plans work today

Section titled “How plans work today”

Every workspace (tenant) has a plan tier. The tier sets resource quotas and a monthly language-model spend cap; it is not a per-request price. The numbers below are the limits the platform enforces on each tier right now. They are reconciled against the code and may be tuned as the platform matures; if you need a higher ceiling, contact the team.

DEV and PRO

Section titled “DEV and PRO”
LimitDEVPROWhere it applies
Knowledge bases110Knowledge Base (RAG)
Indexed chunks (across all your knowledge bases)50020,000the same
Knowledge base storage10 MB500 MBthe same
Monthly language-model spend cap$5$200LLM API: monthly spend cap
Scheduled workflows125Scheduled workflows
  • DEV is for development, prototyping and evaluation.
  • PRO is for production workloads at moderate scale.
  • ENTERPRISE is not self-serve and is not currently assigned to workspaces. If you need higher ceilings, contact the team and we will talk it through case by case. A dedicated database per workspace is designed but not currently offered.

A limit is checked when you create the thing it counts (the next knowledge base, the next upload, the next schedule), and a refusal says which limit you hit and what you have. The spend cap applies for the rest of the calendar month. It is checked before each chat turn, workflow task (scheduled runs included), direct language-model API call (generateText and startTextGeneration) and playground call, and new work is refused with a clear error until the month rolls over or your plan changes. It is a check made before work starts, not an exact meter: work already under way can carry spend past it. Tool-call usage does not count toward it. Speech does: speech-to-text and text-to-speech through the speech service are counted at their underlying cost, and so is the dashboard playground’s speech-to-text, which before 2026-10-21 was not counted at all. A transcript is counted once, at the rate of the engine that produced it (see Speech engines). LLM API: monthly spend cap shows the error the direct API returns.

Reading your usage

Section titled “Reading your usage”

Wetel reports usage; it does not bill it yet. apiUsageSummary returns, for a calendar month, your total language-model cost in USD, the cap you are measured against, a per-model and per-operation breakdown, and a credits figure.

A credit is a unit of usage: one credit is a ten-thousandth of a US dollar of underlying cost (0.0001 USD), with a platform margin applied for speech and avatar usage. For language-model work the credits figure is an approximation (a blended conversion of the cost), and for speech, tool calls and the platform fee it is calculated from fixed per-unit rates rather than estimated; the usage buckets explain which is which. Treat credits as a way to compare and plan, not as an invoice.

What is not built yet

Section titled “What is not built yet”
  • There is no public rate card. We do not publish a per-call or per-token price.
  • There is no usage-based billing, wallet or ledger. Nothing meters your usage against a payment, and you cannot pre-purchase credits.
  • There is no self-serve plan change or checkout. To move tier, or to ask about limits for your use case, contact the team.
Section titled “Related”