Featurebase Pricing Explained: Per-Seat Plans & the $0.49 Fibi AI Fee

If you've got the Featurebase pricing page open in another tab, the seat rows are only part of what you'll pay. The AI charge underneath them moves with your ticket volume.

On 18 June 2026, Featurebase posted a changelog entry called "Fibi 3.0: Faster performance & upcoming pricing update". Two sentences in that entry decide most of your invoice:

"Starting July 1, 2026, Fibi AI resolutions will move from $0.29 to $0.49 per resolution for all new and existing customers"

"Fibi's resolution rate has increased from 50% to 65% across all customer workspaces."

I'd read those two lines together. The unit price went up 69%, and a higher resolution rate means more billable units, so both halves of that announcement push your bill the same way.

Several third-party round-ups still quote the old $0.29 figure, and the AI answers built on them repeat it. Two sources carry today's number: Featurebase's live pricing page and the pricing article in its own help center, both at $0.49.

I'm Mike, co-founder of My AskAI. We help 200+ ecommerce and SaaS businesses run AI customer service inside the helpdesk they already have.

Most weeks I've got per-resolution and per-ticket models side by side in a spreadsheet, working out what each one does to a support budget as volume climbs.

I break down every line item on a Featurebase invoice, run the math at three real volumes, list the costs the pricing page leaves out, and pull together what the public record says about the bill. Then I put the AI charge next to Intercom Fin, Zendesk AI and My AskAI at the same volume, with no "contact sales" answers.

How does Featurebase pricing actually work?

Featurebase runs on two meters. The first is a seat: a flat per-agent subscription that unlocks the helpdesk, the feedback boards, roadmaps, the changelog, the help center and outbound.

The second is Fibi, metered per AI resolution. It climbs with your ticket volume, while the seat line climbs with headcount.

The subscription is all-in on the module side. The pricing page carries a Helpdesk / Feedback toggle, and the per-seat prices are identical whichever way you flip it, so you're buying the whole platform either way.

That's unusual in a category where the feedback tool and the helpdesk are normally two invoices from two vendors.

Fibi sits outside that. Every paid tier gets it at the same rate, and the Free plan doesn't support it at all, so the cheapest way in can't show you what the AI does.

The formula is seats × tier price × 12, plus resolutions × $0.49, plus whichever add-ons you switch on. The two variables you supply are headcount and monthly ticket volume.

What does a Featurebase seat cost, and what's gated behind each tier?

The Free plan is more usable than most free tiers I've seen in this category. You get one seat, live chat, unlimited conversations, a unified inbox and ticketing, the mobile app, assign and snooze, internal notes, one feedback board, one roadmap and 50 help center articles.

SOC 2 and GDPR are included on it, which is rare on a $0 plan.

Free has no AI at all: no Fibi agent, no AI replies, no Copilot.

Growth at $29 turns the AI on: Fibi, email support, AI Copilot, AI replies and macros, custom domains, basic analytics, integrations, five boards and five roadmaps. It's the cheapest seat that can run an AI agent at all.

Professional at $59 is where most support teams I speak to end up, and the feature gates put them there more often than the price does.

API, MCP and webhooks all start at this tier. So do workflows and automations, SLAs, user segmentation, multilingual, team inboxes, Fibi over email, custom AI actions, and the 20 free Lite seats.

Enterprise at $99 is quote-led ("book a demo") and adds multi-brand workspaces, admin SSO with identity management, custom admin roles, the advanced HubSpot integration, Azure DevOps and AD, custom invoicing and 50 free Lite seats.

Admin SSO (SAML or OIDC) and SCIM are Enterprise-only and carry a five-seat minimum, so a two-person team that needs SCIM still pays for five seats.

Multi-brand is Enterprise-only too. And the Slack, Discord, WhatsApp and Microsoft Teams channels, along with advanced analytics, are marked "coming soon", so I'd treat them as roadmap items for now.

I'd read the seat definition before you count heads. Featurebase's seats article says that "Each teammate who needs to access the admin dashboard requires a seat", and a full seat is the billed one with access to everything.

Lite seats are free, and they're built for cross-functional collaborators: sales, marketing, legal, engineering, design.

A Lite seat can view conversations it's mentioned in, submit feedback on a customer's behalf and read feedback and survey responses. It cannot reply to customers, manage settings, configure products or do admin work.

Everyone who answers a customer needs a full paid seat, so that's the headcount I'd build the budget on.

The billing mechanics are forgiving on mid-cycle changes. Featurebase says invoices "always show accurate prorated billing based on the actual days each seat was active", so you're not paying a full month for a seat you added on the 28th.

What does a Featurebase AI resolution cost, and what counts as one?

Fibi is $0.49 per resolution on Growth, Professional and Enterprise. Featurebase's own pricing article puts it in one sentence: "Fibi AI Support Agent is priced at $0.49 per resolution on all plans."

That $0.49 is a recent number. The Fibi 3.0 changelog entry dated 18 June 2026 announced that "Starting July 1, 2026, Fibi AI resolutions will move from $0.29 to $0.49 per resolution for all new and existing customers".

That's a 69% increase, applied to existing accounts as well as new ones, and published in a product changelog. Every page still quoting $0.29 is working from information that predates July, so I'd discount anything built on it.

The same entry reported that "Fibi's resolution rate has increased from 50% to 65% across all customer workspaces", alongside responses roughly 4x faster on complex queries and 2x faster on regular ones.

I ran those two changes through a 10,000-conversation month with the inbound volume held flat. Before that release, a 50% resolution rate at $0.29 billed 5,000 resolutions for $1,450.

The price change alone takes the same 5,000 resolutions to $2,450. The improved rate then adds 1,500 more billable resolutions on top, landing at 6,500 resolutions and $3,185. The AI line came out at 2.2x.

The price rise on its own is +$1,000 a month, or +$12,000 a year, for identical output. The improvement adds a further +$735 a month ($8,820 a year), because 1,500 more conversations now resolve and each one bills.

Shipping a better AI and charging more for it is a defensible commercial decision. Under a per-resolution model, "the AI got better" and "the price went up" both move your invoice the same way, and you absorb both.

Per-resolution pricing bills you hardest when the AI is working best.

Every point of automation you gain converts straight into another line on the invoice. That's a strange incentive to sign up to. The response I see most is teams rationing what they point the AI at.

At 10,000 conversations a month, moving from a 50% resolution rate to a 75% one adds $1,225 a month, or $14,700 a year.

The only thing that changed is the AI.

The 75% comes from our own AI resolution rate benchmarks, which put the median AI-handling rate at 70% across 195 rated deployments, with a 25th percentile of 56% and a 75th of 80%.

It's a directional aggregate across many vendors and many setups, and it skews toward teams confident enough to report a number, so treat it as a field reference point for your own modeling. Featurebase's own 65% fleet figure sits close to the middle of it.

The pricing page can't answer the question I get asked most on calls: what counts as a resolution.

Featurebase frames it as an outcome rather than a billing rule, saying you "only pay when Fibi achieves the outcome you and your customers care about" and describing those as resolved support conversations. The pricing page carries the question "What counts as a Fibi AI resolution in Featurebase?" as an FAQ heading.

Every page that could carry the enumerated version stops short of it: the pricing page, the pricing help article and the Fibi AI Agent overview, which covers capability and knowledge sources without touching billing.

The four rules I'd want defined are all absent: what does and doesn't count, a quiet period, a reopen rule, and whether an escalated conversation still bills. Zendesk publishes a 72-hour quiet period. HubSpot publishes a 72-hour non-escalation rule.

Featurebase publishes the outcome framing and stops there. Until those rules exist in writing, I can't tell you what a reopened or escalated conversation does to your bill.

Fibi and AI Copilot are different products on different meters.

Fibi is the customer-facing agent at $0.49 a resolution. Copilot is agent assist, free for the first 10 conversations per seat each month and $19 per seat per month for unlimited use after that.

Conflating them is the most common error I see in the round-ups.

Fibi reads your Featurebase workspace: help center articles, customer requests, changelogs, roadmaps and past conversations. Featurebase's own claim for it is to "Handle 100% of incoming volume 24/7 using Fibi across all support channels".

What do the other Featurebase add-ons cost?

Four line items sit outside the seat and the resolution fee. They're the ones I see teams discover in month two, and none of them are unique to Featurebase.

The two allowance-based lines are the awkward ones for budgeting. Featurebase publishes the free allowance for both and describes the overage as tiered or usage-based without printing the tiers.

You can model the first 5,000 outbound emails and the first 250,000 translated characters exactly. Everything past that is a conversation with support.

I'd watch Copilot at scale, because it's a per-seat charge once you pass the free allowance. A 20-agent team on unlimited Copilot adds $380 a month, which is $4,560 a year, on top of everything else.

The branding removal is smaller but permanent at $69 a month, or $828 a year, unless you're on Enterprise.

What does Featurebase cost at 1,000, 10,000 and 50,000 conversations a month?

Here are my assumptions, so you can swap your own numbers in. Tier is Professional at $59 per seat per month billed yearly, because that's where the API, workflows, SLAs, multilingual and Lite seats start.

Seat counts are 5, 20 and 50 against 1,000, 10,000 and 50,000 inbound support conversations a month. Fibi is $0.49 per resolution.

For the resolution rate, the headline scenario uses 65%, which is Featurebase's own published fleet average from the Fibi 3.0 entry. Modeling a vendor on the number it publishes about itself keeps the scenario grounded, and it lands close to the 70% field median in our benchmark data.

A conservative 50% floor and a 75% upper case bracket it either side.

Add-ons sit outside these numbers.

At 1,000 conversations a month (small team)

At this volume the seats and the AI are roughly balanced: $295 in seats against $318.50 of Fibi at the 65% rate, for $613.50 a month or $7,362 a year. Even at a conservative 50% the bill only drops to $540, because the seat floor doesn't move.

This is the volume where Featurebase looks cheapest relative to what it replaces, and where $0.49 is a small enough number that I doubt most teams think hard about it.

At 10,000 conversations a month (mid-market)

Here the balance tips hard. Fibi is $3,185 of a $4,365 monthly bill, so the AI is about three quarters of what you pay and the 20 seats are the smaller line.

Play video

Annualized that's $52,380. The effective cost per resolved conversation improves to $0.67, because the fixed seat cost spreads over more resolutions, while the invoice itself keeps growing.

At 50,000 conversations a month (high volume)

At the top end I'd stop talking about seats altogether. Fibi is $15,925 of an $18,875 monthly bill, about 84% of it, and the annual figure is $226,500.

The effective cost falls again to $0.58 per resolved conversation. The check you write is still much larger than at 10,000.

Dropping to Growth at 1,000 conversations and 5 seats takes the bill to $512.50 a month ($6,150 a year), but Growth has no API, MCP or webhooks, no workflows or SLAs, no multilingual and no Lite seats, so for most support teams it isn't the same product.

Going up to Enterprise at 50,000 conversations and 50 seats takes it to $23,325 a month ($279,900 a year).

Here's the full stack once the add-ons are switched on, at 10,000 conversations and 20 seats with a 75% resolution rate:

That's $63,648 a year, or $0.71 per resolved conversation all-in, before any outbound-email or translation overage.

How do I model my own Featurebase bill?

All of the above runs on my assumptions. Yours will differ on seat count, tier, resolution rate and which add-ons you switch on, so here is the same calculation as a prompt you can run yourself.

It walks the three layers in the order they hit the invoice: the per-seat subscription, the Fibi charge, and the add-ons. Two things Featurebase doesn't publish (the overage tiers, and a billing definition of a resolution) I've told it to leave blank and flag for a sales call.

You are helping me model my monthly Featurebase bill. My inputs:

- Support conversations per month: [your monthly inbound volume]
- Agents who reply to customers: [headcount — each one needs a full paid seat]
- Tier: [Growth $29 / Professional $59 / Enterprise $99 per seat per month, billed yearly]
- Billing: [annual, or monthly at $37 / $75 / $129 per seat instead]
- Add-ons switched on: [AI Copilot unlimited at $19/seat/mo? Remove branding at $69/mo?]

Work through it in this order and show the arithmetic at every step:
1. Seat cost = agents × tier price (use the monthly rate if I said monthly).
2. Billable Fibi resolutions = conversations × resolution rate. Model three cases:
   50%, Featurebase's own published 65%, and 75%.
3. Fibi cost = resolutions × $0.49.
4. Add-ons: Copilot = seats × $19 if switched on; branding removal = $69 flat.
5. Monthly total, annual total, and effective cost per resolved conversation.

Then show me how the total moves across the three resolution-rate cases, and what
share of the bill is the AI line versus the seats in each one.

Two things you cannot calculate, so flag them instead of guessing: Featurebase does
not publish overage rates for outbound email above 5,000/month or AI translation
above 250,000 characters/month, and it does not publish an enumerated definition of
a billable resolution (no quiet period, reopen rule or escalation rule). Write
"not published, ask Featurebase" for both rather than estimating a number.

What the Featurebase pricing page doesn't tell you

Most of this is published, just not on the page you're comparing tiers on.

Hidden cost 1: every Featurebase agent needs a full paid seat

Lite seats are free, and they don't reduce the number of paid agent seats you need. Featurebase's seats article is explicit that a Lite seat can't reply to customers, so anyone answering a customer or touching the admin dashboard needs a full seat.

I'd do the multiplication before you commit: a 10-seat Professional team is $59 × 10 × 12, or $7,080 a year in seats, before a single resolution is billed.

Hidden cost 2: the Featurebase AI line climbs as the AI gets better

Featurebase has already demonstrated both halves of this one. The unit count is your automation rate, so improving the AI adds $1,225 a month at 10,000 conversations for a move from 50% to 75%.

The unit price moves too. The Fibi 3.0 changelog took it from $0.29 to $0.49 for "all new and existing customers" from 1 July 2026, in the same entry that announced the fleet rate climbing from 50% to 65%. Together those lifted a 10,000-conversation Fibi line from $1,450 to $3,185 a month.

If you're budgeting a multi-year rollout, I'd price that precedent in: the increase applied mid-relationship to existing customers, and the announcement arrived in a product changelog.

Hidden cost 3: Featurebase's feature gates jump you a tier regardless of headcount

Your requirements decide your tier as much as your team size does, so I'd map the feature list before the headcount. Need the API, MCP, webhooks, workflows, SLAs or multilingual and you're on Professional however few agents you have.

Need admin SSO or SCIM and you're on Enterprise with a five-seat minimum, so a two-person team with a SAML requirement pays 5 × $99, which is $495 a month or $5,940 a year.

Same goes for channels. Slack, Discord, WhatsApp and Microsoft Teams are marked "coming soon", so if one of those is on your requirements list it isn't in the price yet.

Hidden cost 4: the small lines add up

Each of these is a rounding error on its own. At 10,000 conversations and 20 seats they take the bill from $4,855 to $5,304 a month (the 75% scenario, with the add-ons switched on).

Unlimited Copilot is $380 a month for 20 seats, branding removal is $69 a month, and outbound email past 5,000 sends and AI translation past 250,000 characters both meter on rates that aren't published.

Hidden cost 5: you can't model the Featurebase AI unit before you commit

The largest line on your invoice is the one you can't predict from public information. I'd put that question to a Featurebase salesperson before you sign anything.

Featurebase publishes no enumerated definition of a billable resolution, no quiet period, no reopen rule and no escalation rule. You can model your seat cost to the cent, and the AI charge stays an estimate.

What real Featurebase customers say about the invoice

Featurebase is a fully bootstrapped, profitable company founded at the end of 2020 and based in Tallinn, Estonia, with 1,500+ customers across 186+ countries and more than 30m customer conversations handled (those are their published figures). Named users on its homepage include OpenSea, Lovable, Elementor, Instantly and Rize.

The support half of the product is roughly a year old, and Fibi is newer still.

Add a customer base weighted toward early-stage startups on a staged discount, whose invoices look nothing like list price, and you get a base that hasn't been billed for AI resolutions long enough to have generated much public commentary about it.

The commentary that does exist is positive, and it's about price. Two lines from Featurebase's Product Hunt profile, both from around three years ago:

"We moved over from Canny due to price and haven't looked back." Robert Durrette (@ridgeline), via Product Hunt.

"Absolutely fantastic alternative to pricey and bloated competitors in the market." Craig Sartor (@craig_sartor), via Product Hunt.

Both of those predate the helpdesk and Fibi entirely, so I'd read them as an endorsement of the feedback tool's value against Canny. Today's AI pricing is a separate question.

Featurebase's reputation was built on undercutting expensive feedback tools. The AI resolution fee is a much younger line, and a far more volume-sensitive one that the existing customer base is only now starting to feel.

Anyone buying Featurebase for the AI today is in the first cohort to be billed at $0.49.

Does Featurebase have a free trial, and what are the discounts and terms?

There are two ways in without a card, and the first is the Free plan: $0 forever with one seat, live chat, unlimited conversations, a unified inbox and ticketing, one board, one roadmap and 50 help center articles.

Separately, Featurebase's help center confirms a trial of the paid features, saying "You can get started with a Free plan or trial - no credit card required!".

Because Free doesn't support AI, the free-forever route can't evaluate Fibi, so I'd take the trial if the AI is the thing you're actually assessing. If the AI is the thing you're buying, the paid-feature trial is the route I'd take.

The pricing page defaults to yearly billing and each plan card carries a toggle for monthly. Monthly runs roughly 27% more: $37, $75 and $129 a seat against $29, $59 and $99.

The early-stage startup program carries the largest discount in the category.

The pricing page markets it as "86% off" plus "1 year of Fibi AI free". The program page defines the AI benefit as 50 free Fibi resolutions per month in year one, with additional resolutions at $0.49 each.

The seat discount is large: $450 a month down to $63 for six Professional seats, which works out at $10.50 per seat per month in year one.

The AI benefit is smaller than the headline suggests, and I'd run your own volume through it first.

Take a startup on the deal doing 1,000 conversations a month at a 75% rate on 5 seats: $52.50 in seats plus 700 chargeable resolutions at $0.49, which is $343, for $395.50 a month.

Even in the "free AI year", 86.7% of that bill is the Fibi line, because 50 free resolutions covers 6.7% of the month's total.

Eligibility is tight: founded less than 2 years ago, fewer than 6 employees, less than $3M raised, never previously on a paid plan, an application within 90 days of signup, and a 6-seat cap. In return you get Professional-plan features.

How does Featurebase pricing compare to the alternatives at the same volume?

Every row below is the AI charge only, at 10,000 conversations a month with a 75% resolution rate.

For My AskAI, Intercom Fin and Zendesk AI those sit on top of the helpdesk seats you keep paying either way. Featurebase's seats buy the helpdesk itself, so I've counted them separately.

And the platform line, kept separate so nothing gets double-counted:

Featurebase does well against the native helpdesk AIs. At $0.49 it undercuts Intercom Fin's $0.99 by half and Zendesk AI's ~$1.50 by two thirds.

Against our flat per-ticket model the direction is the other way, and the gap moves with your resolution rate. At 50% the Fibi line is about 1.9x My AskAI's, and at 75% it's about 2.8x. RecruitCRM turned down Fin at $0.99 because that math didn't work for them.

Featurebase and My AskAI sit in different categories.

Featurebase is the helpdesk, plus a feedback and roadmap suite. We're an AI layer that runs inside Zendesk, Intercom, Freshdesk, Gorgias or HubSpot.

Adopting Fibi means moving your support onto Featurebase.

Where Featurebase wins outright is consolidation. One per-seat price covers helpdesk, feedback boards, roadmap, changelog, help center and outbound, which is four or five tools' worth of subscriptions for teams currently buying them separately.

Its startup ladder runs for three years. It's EU-built and EU-hosted.

And SOC 2 Type II plus GDPR land on every tier including Free, which is the detail I'd put in front of a security reviewer first, because most vendors gate it behind Enterprise.

Feedback tools are the spend Featurebase displaces.

Canny is free up to 25 tracked users and $79 a month on Pro at 100+ tracked users, with the meter running on tracked users (anyone who posts, votes or comments).

Productboard is $19 per maker per month on Plus and $59 per maker per month on Business, billed annually. Featurebase folds that whole category into the same $59 seat as the helpdesk.

If you're consolidating tools, you're early-stage enough for the discount ladder, and your AI volume is modest, Featurebase is a strong buy.

Is Featurebase actually worth the money?

On the security question, Featurebase's security FAQ states it's "officially SOC 2 certified" with the Type II report available on request, "fully GDPR compliant", and hosting data in the Netherlands, Germany and Ireland.

There's a published data processing agreement too. For an EU buyer that's a strong position at every price point, Free included.

The savings calculator on Featurebase's pricing page defaults to 4 seats and 500 AI resolutions against "an equivalent Zendesk plan" and reports "you save $10,668" a year.

That's Featurebase's own comparison on Featurebase's own assumptions, and it models one point on the curve. Swap in your seat count and your resolution volume and the number moves.

With Featurebase you move your support onto its platform and get the whole suite in one per-seat price.

My AskAI asks you to change nothing, and we charge about $0.10 a ticket whether the AI resolves it or not. Both are reasonable buys at 1,000 conversations a month. The gap opens at 10,000 and above.

FAQs