Pricing
A plan grants credits each month. Every pass spends them at a rate you can read before you press anything, and nothing here bills by the seat.
Free
No card
One fair a month, free. Any of the thousands already on file, up to 1,000 stands, read in full with a profile for every company on it. Plus 3,000 credits a month for scoring.
Researching a fair nobody has read yet needs a plan
Or pay yearly and get 2 months free.
Starter
9,500 credits a month
One seller, a couple of fairs a quarter.
- The whole catalog, browsable
- Exhibitor lists with company profiles
- ICP scoring on every booth
- Named people, the agenda, and their LinkedIn
- Company Search and People Search
- Ask AI beside every list
- CSV export
- MCP server
Cancel any time · unused allowance expires monthly
Pro
24,000 credits a month
A team working a season of fairs.
- Push companies and people to HubSpot or Attio
- The REST API and webhooks
- 300 API requests an hour, 3,000 a day
Everything in Starter, plus the above.
Cancel any time · unused allowance expires monthly
Scale
74,000 credits a month
Several sellers, several halls, every month.
- 600 API requests an hour, 9,000 a day
- A higher daily spend ceiling
Everything in Starter, plus the above.
Cancel any time · unused allowance expires monthly
What a credit buys
A credit is a unit of work rather than a subscription tier, and one dollar is 250 of them. A fair is priced by its size, so these are rates rather than totals.
A fair, in full
1,500–35,000 creditsOne press: every exhibitor, and what each of them sells. Priced by how big the fair is: 1,500 up to 100 stands, 2,500 to 250, 4,000 to 500, 7,500 to 1,000, 18,000 to 2,500, 35,000 beyond. If it turns out bigger than we priced it, we read it all anyway; if it turns out smaller, the difference comes back.
ICP scoring
250 credits a fairEvery booth judged against who you sell to, with the reason quoted.
People
1,300 + 1,000 / 40 standsEvery stand walked for its contacts, plus four runs over the organiser's programme, which brings back the timetable as well as the speakers, on one purchase rather than two. A 200-stand fair is about 6,400 credits. Never a constructed email address.
Company Search
220 creditsOne company: which fairs it will be at, who is going, what it posted.
People Search
220 creditsOne person: where they will be, what they said, and their profile. Asking who somebody works for is free.
Ask
24 creditsOne question against what your account already holds. It reads; it never buys.
Find fairs
510 creditsA sweep of your region and vertical for fairs worth the flight, each one scored against who you sell to. Nothing about a fair is bought yet. This finds them and ranks them.
Find matches
250 creditsEvery company the catalog already holds, ranked against your ICP with the sentence it was judged on. Searching wider, meaning paying to go and find more of them, is 510.
A fair, priced end to end
Pengo quotes every press before you make it. This is what the quote looks like on a middling fair.
A 200-stand fair, read in full
- The fair, every exhibitor and what they sell
- 2,500
- ICP scoring
- 250
- Total
- 2,750 credits
People are the separate purchase and still the larger one: every stand walked, plus the organiser's programme, is about 6,400 credits at this size.
Which means the plain version: Starter covers 3 fairs of that size in a month, Pro covers 8 and Scale covers 26. Walking the people at one of them is roughly another two.
Top-up packs
A plan's allowance expires each month. A pack carries over, sitting on the account for twelve months from the day it is bought.
Packs are priced slightly above the plan rate rather than below it. A top-up buyer has just finished a month’s allowance, which means fresh fairs and the most expensive research we run. A volume discount there would put the cheapest credits on the costliest work.
The API
Read-only, versioned at /api/v1, and authenticated with a key you mint in settings. It is what Zapier, n8n and Make connect to, and it starts at Pro.
Pro
300 an hour · 3,000 a dayAcross every key on the account, and no single key may draw more than 150 of the hour, so one integration left in a retry loop cannot take the others down with it.
Scale
600 an hour · 9,000 a dayAcross every key on the account, and no single key may draw more than 300 of the hour, so one integration left in a retry loop cannot take the others down with it.
Page size
100 rowsEvery list answers a cursor. A two-thousand-stand fair is twenty requests rather than one, and the response says how many rows are left.
What it will not do
- It spends nothing. Every purchase stays a press in the app, against a stated price.
- It starts no research. The endpoints read what has already been paid for.
- It writes nothing, anywhere. Every method but GET is refused.
- It reads one account: yours. Nothing another customer bought is reachable through it.
The rest of it, without the asterisks
- Free is a tier rather than a trial, so nothing starts billing you quietly. It is one fair a month from what the catalog already holds, plus credits for scoring, and it does not ask for a card.
- Cancel and you land back on free rather than locked out. Credits bought in a pack keep the rest of their twelve months.
- Your monthly allowance is spent before anything you bought outright, so the expiring balance goes first.
- The quoted price is the price charged. A pass that establishes nothing settles as failed and releases in full.
- Looking up a fair that is not on file yet costs nothing. Type its name, and a couple of minutes later it is in the catalog with its dates, its size and its price.
- The first read of a fair takes fifteen or twenty minutes, and the stands land as they are read. A fair only ever has to be read once, so a fair already in the catalog opens instantly.
- A fair somebody else already researched costs you the same as it cost them, and arrives with no wait at all. The catalog is shared; your scores are not.
- Every plan includes the MCP server. The REST API and webhooks start at Pro, at 300 requests an hour.
Questions people ask before they pay
The awkward ones included. A pricing page that answers only the comfortable half is a refund in week one.
Pengo is trade-fair research for people who sell at them. It reads a fair's exhibitor directory end to end, establishes what every company on it sells, scores each booth against who you sell to, and names the speakers and stand contacts, so you arrive with a walk route instead of a floor plan.
It scores every stand against a description of who you sell to, and quotes the sentence it judged on. A score with no reason attached is a number you have to trust; a score with the company's own words under it is one you can check in about four seconds and overrule.
From the fair's published directory and the companies' own pages, read at the time you ask. Every fact carries the page it came from. Pengo does not resell a scraped list: a directory is read to its end, and a fair whose list is only partly published says so rather than passing a fragment off as the whole floor.
Fifteen to twenty minutes for a fair nobody has read yet, and longer for a very large floor. The agents work the directory page by page and the stands land as they are read rather than all at once, so the list fills while you are reading it. A fair already in the catalog is on screen the moment you press.
Type its name and Pengo goes and finds it. That costs nothing and takes a couple of minutes. The fair lands with its dates, its venue and its size, priced and ready to unlock. The catalog is not a fixed list of events: it is everything anybody has ever looked up, and it grows every time somebody looks one up.
Only where a published page actually states a business email address. Pengo never constructs one from a person's name and company domain, and its MCP attendee tool does not return email addresses. A guessed address is indistinguishable from a real one to whoever sends to it, so guesses are discarded.
A fair is priced by its size, and every press is quoted before you make it. A 200-stand fair, read in full, meaning the exhibitor list, a profile for all 200 companies, and ICP scoring across the floor, costs 2,750 credits, or about $11. People are a separate purchase and the larger one.
There is a free tier rather than a trial, so nothing starts billing you. It gives you one fair a month at no cost, any fair already on file, up to 1,000 stands, read in full, plus 3,000 credits a month for scoring. Researching a fair nobody has read yet needs a plan, from $39 a month.
Yes. Every plan is also sold yearly at 10 months' price, so Starter is $390 a year instead of $468. It is the same plan on a different invoice: 9,500 credits still land every month, and unspent allowance still expires monthly. Switch the pricing table to annual to see all three.
One fair a month at no cost, from everything already researched: every exhibitor on it and what each of them sells, up to 1,000 stands. Plus 3,000 credits a month for ICP scoring, the whole catalog to browse, and your notes and exports. Reading a fair nobody has researched yet, and the API, need a plan.
Yes, on an active plan through Pengo's remote MCP server. Claude, Claude Code and Cursor have direct setup paths; ChatGPT requires a workspace and plan where custom MCP connectors are enabled, sometimes by an administrator. Five tools read, and one updates only your booth route, visit state and notes.
Yes, a read-only REST API from Pro up, plus the MCP server on every active plan. It answers your fairs, exhibitor lists, people and customer-facing credit use, the same rows the app draws, behind a key you mint in settings. REST writes nothing and starts no research: every purchase stays a press in the app.
300 requests an hour and 3,000 a day on Pro, 600 and 9,000 on Scale, across every key on the account. No single key may take more than half the hour, so one integration in a retry loop cannot starve the others. Lists answer 100 rows and a cursor.
You pay the same as they did, and you get the answer immediately. The exhibitor list and the company profiles are shared facts, so the second buyer launches no research. The price does not move, because a bill that depended on whether a stranger wanted the same fair last week is not a price.
It settles as failed and releases the credits in full. A pass that established nothing has nothing to sell you, and reporting zero rows against a charge is the one outcome that makes a product impossible to trust. Rows that fail a check are dropped rather than repaired, and the number dropped is on screen.
Price a fair
before you book the flight
Every press is quoted first, and the quote is what you pay.