Pricing
Credit usage, live account prices, reservation settlement and variable AI inference — with unchanged USD compatibility fields.
buzzabout presents product usage in credits. Data work is priced by results collected or processed; AI inference is priced separately from tokens consumed. Subscription and top-up payments are cash in US dollars. Credits project one canonical USD wallet, not a second balance.
Additive public credit fields
Existing USD fields keep their names and units. For credit presentation,
read prices_credits on GET /v1/prices,
usage.credits_used / usage.credits_total on
GET /v1/me, and quantity_credits on history.
billing.unit selects credits or usd presentation.
Unavailable credit projections are null, never an invented rate or
a free price. The removed GET /v1/credit_prices route does not return.
What an action costs
Ordinary registered plans use the same credit action weights. Read your
account's stored prices from GET /v1/prices
rather than reconstructing a quote from a plan name or assuming a missing
category is free.
| Public category | Credits | Charged work |
|---|---|---|
mention | 1 | Each mention collected by a dataset run or modern listening-agent collection |
audience_profile | 3 | Each profile collected by an audience-dataset run |
post_processing | 0.5 | Each post processed by a custom-parameter or pattern-detection run |
preview_generation | 1 | Each source in a public research preview; URL previews refund sources returning no data |
Other configured actions remain distinct: legacy tracking dispatch costs 15 credits, report generation 300, and legacy audience-insights generation 600. These are not additional run fees on every modern listening-agent or audience-profile collection. Balance sync has no charge.
New V3 plans include the following monthly allowances and usage rates. Seats include the account owner.
| Plan | Monthly price | Annual price | Monthly credits | Credit usage rate | Total seats |
|---|---|---|---|---|---|
| Starter | $98 | $980 | 2,000 | $0.049 | 1 |
| Pro | $178 | $1,780 | 5,000 | $0.0356 | 3 |
| Business | $450 | $4,500 | 20,000 | $0.0225 | 5 |
Onboarding starts with a $1 one-day trial with 1,000 credits. The first paid month is $49 / $89 / $225, then the monthly prices above. Annual plans are paid up front with two months free. They receive the same monthly credit allowance and usage rate as the corresponding monthly plan, without a trial, activation fee or first-month coupon.
Existing V3 subscriptions keep their purchased allowance and usage rate on renewals and top-ups until the subscriber deliberately changes plans. The plan picker shows only current offers; changing plans moves the subscription onto the selected current terms. Older offers cannot be selected again. Existing captured payments retain their captured terms. See the plan comparison.
Changing plans and cancellation
Change plans in Plans & top-ups. An upgrade starts a full new monthly or annual billing period immediately and grants the new plan's full monthly credit allowance. Eligible unused paid subscription value is applied to the new invoice based on credit usage, rather than time remaining. Any value left after the invoice is settled becomes billing credit toward future invoices. The new monthly allowance replaces the old one. The quote shows the timing, amount due, and new credit usage rate before you confirm.
Downgrades take effect at the end of the current billing period. To cancel, use Cancel plan at the bottom of Plans & top-ups. Cancellation takes effect at the end of your paid period, and access continues until then.
AI inference is charged separately
POST /v1/ask and MCP buzzabout__ask are
charged for inference consumed by the turn. The token-based charge
settles after the reply; it is not a fixed 15-credit assistant fee.
Data collection or processing requested by the assistant is charged
separately at its applicable action prices.
MCP buzzabout__query invokes
an AI planner and also incurs variable inference charges, even when
querying already-collected data. Provider costs and serving-model terms
determine the canonical USD charge; customer credits project it at the
account rate. Auto served by Luna uses provider cost × 1.50; explicit
Luna and other serving cases use × 1.10. There is no flat planner fee.
Reading existing rows, saved results, messages, run status, account info
or prices does not start another paid run. Custom-parameter test previews
are free. The app's keyword estimates and Preview mentions inspection
are distinct from a billed public research preview; asking the assistant
to produce or interpret them can still consume charged inference.
Reservation-based charging
The four public data categories reserve enough canonical USD for the upper-bound result count before work starts and refund unused reservation value on completion. In credit terms, a request for 500 one-credit mentions returning 412 reserves 500 credits, settles at 412, and releases 88. The server converts those amounts using the account's captured terms. Failed work releases its unused reservation; already-settled work follows that operation's actual-result rules.
A public research preview reserves
one credit per source. URL previews refund sources returning no data.
MCP collect_mentions uses this billed preview internally: a thin
preview can prevent collection spending without making the preview free.
Token-settled AI inference is separate from result-count reservations.
Insufficient balance
A data request that cannot fund its reservation is rejected before that
work starts. Ask rejects an already-spent balance with
402 insufficient_credits_ask, without starting or charging a turn.
Keep handling the existing insufficient_credits* machine codes.
Existing required / available fields remain USD; read explicitly
named credit fields when provided, never relabel those USD values.
Agent configuration and existing dashboards may remain available, but
new paid work requires available balance. Unlimited labels do not
bypass configured wallet, concurrency or research-depth limits.
Programmatic access
curl https://api.buzzabout.ai/v1/prices \
-H "X-API-Key: $BUZZABOUT_API_KEY"{
"status": "success",
"data": {
"prices_credits": {
"mention": "1.000000",
"audience_profile": "3.000000",
"post_processing": "0.500000",
"preview_generation": "1.000000"
},
"billing": {
"unit": "credits",
"catalog_version": "2026-09-18",
"precision": 6
}
}
}This excerpt shows the credit fields; the response also retains its USD
compatibility fields. prices_credits contains six-decimal strings, or
null for an unavailable category. billing can be null when contract credit terms
are unavailable; USD prices remain readable. currency, rate and
prices retain USD compatibility semantics. The legacy rate can also
be null on a category-priced contract. Keep full decimal precision in
arithmetic and round only for display, never to determine affordability.
Top-ups and history
The app lets you choose credits and shows the cash charge secondarily. Top-ups are paid in exact USD cents at the server-confirmed account rate, with a $5 minimum. Captured payment terms determine the credit receipt; a later plan change does not rewrite it. Cash payments, invoices, affiliate payouts and the USD ledger remain dollar-native.
GET /v1/me/usage_history is an account-wide,
paginated spend-only feed. quantity remains non-negative USD charged.
quantity_credits, credit_rate_usd, billing_catalog_version and
credit_value_basis identify its credit equivalent:
snapshot: complete captured event terms are preserved, even without a current account rate.current: a pre-snapshot event is projected at today's rate, not reconstructed as an event-time credit charge.nullmetadata: an uncaptured event has no available current rate.
Current wallet totals use the active account rate. Do not sum differently
rated historical credit snapshots to compute today's balance. Categories
may include ai_assistant, report, audience_research, tracking,
research_mention, tracking_mention and preview_generation; treat
unknown strings defensively. An ai_assistant history row is not evidence
that current Ask is free.
Next steps
GET /v1/prices— live prices in both units.GET /v1/me— current usage and credit history.- MCP tools — inference versus free raw reads.