Skip to content
Voqalize

Usage and limits

get_usage(tenant, period="") returns one tenant’s counters for one billing period. period is YYYY-MM in UTC; empty means the current month.

These are counters maintained as the sessions happen rather than a scan of your history, so the answer costs the same in a tenant’s fortieth month as in its first, and a period with no sessions returns zeros rather than taking longer to say so.

The usage quantity

Section titled “The usage quantity”

duration_secs comes from Voqalize’s own measurement of the session — the time it actually held, not the span from when a token was minted. It runs from the moment the call connects to the moment it ends. A session that was created and never answered contributes zero.

That span is what a recording is cut to, so the number counted here is the length of the audio you can play back. Silence inside a call is inside both: a caller who stops talking has not ended the session, and the track carries those seconds as silence.

Voqalize is priced per conversation minute, with the developer’s LLM billed separately by its provider. Each plan’s included minutes and rate are on the pricing page.

agents breaks the same numbers down per agent, busiest first.

The gap that tells you something

Section titled “The gap that tells you something”

sessions_created minus sessions_started is the sessions nobody ever joined: a session was minted, and no user ever connected to it.

A few of those are normal — someone opened the page and left. A wide gap is a broken embed rather than a broken agent, and it is the one number here worth an alert. The usual causes are a page that mints on load instead of on click, an origin the pk_ key’s allowlist rejects at the offer, or a connect that throws before the transport starts.

list_sessions shows the sessions behind any number, and reading a call back shows how far each one got.

Limits

Section titled “Limits”

Concurrent sessions, included minutes and storage are set by your plan; the pricing page lists each plan’s allowance. Talk to us before a launch that needs more than a plan carries.

A session lasts at most one hour: the voice tier ends it at the hour with end_reason max_duration. A session still active five minutes past the hour reads lost, because the voice tier never reported how it ended.

The MCP server is rate limited per tenant, which shapes how you poll. A refused request is a 429 with a Retry-After header; wait that long and send it again. A dashboard that fires a burst of reads per page is well inside it; a loop that reads every session every minute is not the shape to build. Read a session when something happened to it, which is what the session_id join is for.

Section titled “Read next”