skip to content
fizzgig
sign in
// suite
architectureknow what you builtoperationsknow it’s runningauditknow it’s safe to shipintegrationspricingdocschangelog
// community
sourdough startersproject kicks-off in 1 prompt
sign inrequest access →
// pricing

pricing that grows with your product.

free while you're prototyping. £15 the moment you connect a real stack. £75 when you're running a real product with a real team. every tier gets all three pillars in full. price scales with your team — use the seat counter below to see your monthly total.

// hobby

hobby

£0forever · solo

for weekend prototypes, trying it out, learning what fizzgig catches that your other tools miss. connect one platform, run 50 AI-editor calls, see the shape. when you hit the cap, you'll know if it's for you.

solo · personal team of 1
1 project · 1 connected platform
50 MCP calls / mo
daily refresh cadence
  • +every audit tool · every dashboard
  • +perfect memory + forward plan + stack-map
  • +daily digest email
  • −inviting collaborators (startup +)
  • −more than one platform (startup +)
  • −faster refresh than daily (startup +)
start hobby →
most founders start here
// startup

startup

£15/ mo · 1 seat
starts at £15 - add teammates below

for founders shipping their product. multiple platforms wired up, decisions captured, costs tracked, deploys watched. add teammates as the team grows.

team size
1
£15 first seat + £5 each additional
500 MCP calls / mo pooled · +200 per extra seat
1 platform included · +£3 each
hourly refresh + on-demand
  • +everything in hobby
  • +invite teammates + role management
  • +anomaly detection + notifications summariser
  • +cost-optimise recommendations
  • +DDQ answerer against live infrastructure
get startup →
// scale-up

scale-up

£75/ mo · 1 seat
starts at £75 - add teammates below

for teams building at production scale. multi-product portfolios, 15-minute refresh so the notifications surface actually feels live, pooled MCP capacity for a full team's editors, and $25 USD/mo of deep-audit credit included so the dual-agent LLM review is basically on-tap. base tier billed in GBP; deep-audit credit denominated in USD because Claude + Gemini bill the underlying LLM calls in USD.

team size
1
£75 first seat + £15 each additional (GBP)
5,000 MCP calls / mo pooled · +1,000 per extra seat
5 platforms included · +£3 each
15-min refresh + on-demand
$25 USD / mo deep-audit credit (bundled)
  • +everything in startup
  • +$25 USD / mo of deep-audit credit — ~5 deep-tier or ~40 quick-tier scans (LLM providers bill USD)
  • +priority queue + early access to new tools
  • +shared project workspaces + team ownership
  • +higher call cap for team-scale usage
get scale-up →
// what you're paying for

two axes. one honest bill.

// 1
platforms - what we watch for you

each connected platform (supabase, vercel, github, stripe, neon, netlify, gitlab, railway, more shipping) runs on a cron every 24h to keep your snapshot fresh - daily on hobby, hourly on startup, 15-min on scale-up. more platforms = more work on our side = the honest place to meter.

// 2
MCP calls - what your AI editor does

every time your AI editor calls a fizzgig tool (memory_record, plan_add, stack_map, rls_checker, etc), that's one MCP call. cron pulls don't count against this - they're covered by the per-platform fee. only calls you actively trigger meter here.

// 3
hit the wall? upgrade prompt

when you hit either cap, the next tool call returns a structured upgrade nudge - surfaced verbatim by your AI editor. soft warnings at 80% + 95% via dashboard banner + email. no surprise bills, no pay-as-you-go gotcha.

// 4
we never see your code

MCP calls carry file paths + parameters, not source bodies. platform snapshots pull public metadata (deploy counts, advisor warnings, cost estimates) not application code. stack-map derives from tool outputs. privacy is the default, not a configuration.

// refresh cadence

how live is live?

the notifications surface, cost tracker, and anomaly detector are only as fresh as your last snapshot pull. tier picks the cadence.

// hobby
daily

fine for weekend projects - you'll see yesterday's advisors, deploys, and cost warnings

// startup
hourly + on-demand

the sweet spot for shipping a product - advisors fresh within an hour, plus refresh anytime

// scale-up
15-min + on-demand

team-scale - notifications feel real-time, anomaly detector catches issues before your users do

// savings

two ways to pay less.

// annual

12 months for the price of 10.

pay annually and get ~17% off - that's two months free on any tier. cancel anytime, refund prorated.

// student

30% off for students.

verified .edu / .ac.uk email gets 30% off any paid tier. stacks with annual. coupon applied automatically at checkout. for actual students - not a coupon-farming exercise.

// faq

questions you might have.

what counts as an MCP call?
every time your AI editor calls a fizzgig tool (memory_record, plan_add, stack_map, any audit tool) that's one MCP call. platform cron pulls DON'T count - they're covered by the per-platform fee. only calls you actively trigger meter against your monthly cap.
what counts as a platform?
a connected credential in /dashboard/integrations. supabase project, vercel project, github repo, stripe account - each = one platform. one PAT can serve multiple projects at fizzgig's side, but each distinct credential row = one platform on the meter.
what happens when i hit the cap?
the next tool call returns a structured upgrade prompt your AI editor surfaces verbatim. no surprise bills. soft warnings at 80% + 95% via dashboard banner + email so you see it coming.
why platform-based pricing?
because it's the honest cost driver on our side. each platform = a cron running every 24h/1h/15min + snapshot storage + downstream ai-summariser calls when signals fire. one credential = a real ongoing service, priced accordingly at £3/mo - less than a coffee.
can i use fizzgig with claude code?
yes. it works in any mcp-compatible client: cursor, claude code, windsurf, bolt, replit. any editor that speaks MCP as a client works - the config is the same URL + Authorization header.
do you see my code?
no. MCP calls carry file paths + parameters, not source bodies. platform snapshots pull public metadata (deploy counts, advisor warnings, cost estimates) not application code. stack-map derives from tool outputs.
can i mix and match?
the tier picks your baseline (calls cap, platforms included, refresh frequency). extra platforms beyond the tier's included count are +£3 each. no per-user seats, no add-on modules - the tier is the whole product.
is there a self-hosted / enterprise tier?
not yet. enterprise (SSO, SLA, custom retention, self-hosted control plane) ships when someone asks - the answer is usually "yes, let's scope it." right now scale-up is the top self-serve tier.
what about pay-as-you-go top-ups?
deferred until we see how many users hit their cap mid-month. for launch: hard cap with upgrade prompt, no overage billing. simpler model, no surprise bills.

ready to connect your first platform?

no credit card to start hobby. move up when you outgrow the cap - no data migration, no rewriting.

start hobby - free →
fizzgig

the fluffy guardian of vibe-coded products. growls at insecure code so you don't have to.

all systems operational

suite

architectureoperationsauditintegrationspricingdocschangelog

community

sourdough startersdiscord (soon)github (soon)x / twitter (soon)rss (soon)

company

aboutcontacttermsprivacy
© 2026 fizzgig.
v1.5.0 · 2026-07-26