v0.1· early preview · open for design partners·get in touch →

the back-office your api
product is missing.

auth, queryable logs, resource-keyed journeys, event workflows. type dep_K7sQpL2mXr9 and see every operation that built it.

no card · 100k req free

node.jspythongorustcopy

// wrap your handler. ship.
import { apikite } from '@apikite/sdk'

const ak = apikite({ adminKey: process.env.APIKITE_ADMIN_KEY })
export const POST = ak.protect(handler)

resource explorer // production

orca-labs · today

resource:dep_K7sQpL2mXr96 events · 3m 24s

14:32:01

// customer's first ever api call

POST/v1/deploys201

created · ref main@a3f9c8d · project proj_8aJ4qV2N · workflow first_call fired

14:32:14

GET/v1/deploys/dep_K7sQpL2mXr9200

read · status building

14:35:01

PUT/v1/deploys/dep_K7sQpL2mXr9200

updated · attached cfg_y3mR9xKp · diff 2 vars

14:35:18

POST/v1/deploys/dep_K7sQpL2mXr9/promote503

error · upstream timeout · idempotency-key idem_p3o2k… · workflow 5xx-burst fired

14:35:24

POST/v1/deploys/dep_K7sQpL2mXr9/promote200

promoted · environment production · idempotency-key idem_p3o2k… matched · status live

14:35:25

ACTworkflow:notify-release → #releases-prod200

posted to slack · workflow deploy-success fired · channel #releases-prod

engineered to handle billions of requests / month

◆ ai inference · 1.2b req/mo▲ ci platform · 800m req/mo◯ dev tooling · 250m req/mo▣ mobile saas · 320m req/mo※ iot telemetry · 600m req/mo◇ data platform · 180m req/mo

№ 01auth · keysmiddleware live

one line in every framework.
zero spaghetti.

wrap your handler with ak.protect in hono, next.js, or anything that speaks web request/response. keys verified, scopes checked, environments enforced — with a real reason in the body, not just a status code.

1import express from 'express'

2import { apikiteExpress } from '@apikite/sdk/express'

3

4const app = express()

5const ak = apikiteExpress({ adminKey: process.env.APIKITE_ADMIN_KEY })

6

7// verifies the bearer key, checks scope, captures the request.

8// 401 / 403 with a real reason in the body — not just a code.

9app.use('/v1', ak.protect())

npm i @apikite/sdkserver.js9 linesjavascript

api keys// 4 active

verifying

production· web

live

sk_live_••••••••••••f29acopy

read:write↑ 2,481 / day

mobile· iOS

live

sk_live_••••••••••••b401copy

read↑ 1,142 / day

edge· workers

live

sk_live_••••••••••••3d72copy

read:write↑ 812 / day

ci· staging

test

sk_test_••••••••••••0c8ecopy

read:write↑ 42 / day

verifications// every request, every key, every verdict

3.8 req/s

14:32:01GET/v1/users/u_M9pQ4kLsk · f29a20028ms

14:31:54PUT/v1/projects/proj_8aJ4qV2Nsk · b4014036ms

14:31:43GET/v1/projects?limit=20sk · 3d7220056ms

№ 02journeysresource correlation

type any id
see every operation
that touched it.

no more grepping logs by timestamp and guessing. drop a deploy, a user, a project, even a tenant id — apikite returns the chronological story across every request and retry.

orca-labs/production/Journeys/dep_K7sQpL2mXr9

live

resource lookup

id:⌘K

deploy·/v1/deploys/:id

dep_K7sQpL2mXr9

live · production · main@a3f9c8d

events

7

span

6m 42s

errors

1

lifeline// time-compressed

requestinbounderrordeploy

14:32:0014:38:42

events// chronological · req+res payloads captured

14:32:00+0ms

POSTgithub/v1/inbound/github/push20042ms

received · ref refs/heads/main · sha a3f9c8d · pusher orca-bot

14:32:01+1s

POST/v1/deploys201186ms

created · ref main@a3f9c8d · project proj_8aJ4qV2N · trigger github.push

14:32:14+13s

GET/v1/deploys/dep_K7sQpL2mXr920028ms

read · status building

14:35:01+3m

PUT/v1/deploys/dep_K7sQpL2mXr920094ms

updated · attached cfg_y3mR9xKp

14:35:18+17s

POST/v1/deploys/dep_K7sQpL2mXr9/promote200412ms

promoted · environment production

14:37:58+2m 39s

POST/v1/deploys/dep_K7sQpL2mXr9/rollback5035,021ms

error · upstream timeout · workflow 5xx-burst fired

breakdownclick to filter

requests

6

errors

1

relatedpivot →

project

→

proj_8aJ4qV2N

3 events shared

config

→

cfg_y3mR9xKp

2 events shared

tenant

→

tnt_orca_labs

6 events shared

try anotheror paste any id from your logs⌘K

№ 03workflowslive preview

write rules.
watch them fire.

attach triggers to any signal in your api — first calls, 5xx bursts, quota crossings, resource lifecycle events — and run actions against slack, pagerduty, email, or any webhook. no glue code, no log shipping pipeline. point at the rule, deploy.

this hour42 firesacross 4 workflows · 0 failed

request logs// streaming

4.2 req/s

timeverbendpointtenantlat

14:31:54GET/v1/projects?limit=50tn_orca52ms

14:31:48GET/v1/users/u_M9pQ4kLtn_orca34ms

14:31:41PUT/v1/projects/proj_8aJ4qV2Ntn_orca88ms

14:31:34GET/v1/usage/summarytn_orca12ms

14:31:26POST/v1/sessionstn_orca24ms

showing 5 of 49,201open log explorer →

→

workflow// first-call-alert

1,204 runs·0 failed·p99 312ms

whentrigger

eventrequest.first

filtertenant.is_new

cooldown24h per tenant

↓

thenaction

runslack.post

channel#signups

templateonboarding-v2

last fired just nowfor tn_acme-corp

→

slack

connected

#signups

14 members|celebrate every new integration

ak

apikiteapp2:14 pm

first call · helix-systems just made their first request to your api

tenanttn_helix

endpointPOST /v1/auth/exchange

latency168ms

view trace in apikite ↗

ak

apikiteapp2:21 pm

first call · orca-labs just made their first request to your api

tenanttn_orca

endpointPOST /v1/auth/exchange

latency142ms

view trace in apikite ↗

+Message #signups@:↗

trigger source

events from request logs, resource lifecycle, workflows, or cron

action targets

slack · pagerduty · email · webhook

define in

ui · yaml · terraform · lives in your repo, deploys with the rest

№ 04dxtenant onboarding

every tenant's first 200,
measured.

sign-up to first successful call, per tenant, per endpoint. set the targets your tenants should hit. ship a regression, the tenants stalling on it surface before they churn.

tenant dx // orca-labs · last 30d

2,481 tenants·all targets on track

time to hello worldp50

4m 12s

↓ 1m 38skey.created → first 2xx

activation7d

62%

↑ 8pptenants reaching 5+ calls

first-call successany endpoint

84%

↑ 3ppfirst attempt 2xx

retention to call #224h

71%

↑ 1ppall tenants

ttfhw p50

key.created → first 2xx

4m 12s / < 10m

passing

activation 7d

tenants reaching 5+ calls

62% / > 50%

passing

first-call success

first attempt, any endpoint

84% / > 75%

passing

uptime

99.99% target

multi-region failover · zero-downtime deploys

overhead

<8ms p99

benchmarked across all sdks

scale

billions / mo

engineered capacity per region

sdks

14 langs

node · py · go · rust · rb · ex · 8 more

// pricing

free to start. honest at scale.

per-request pricing with no surprise overages. every tier ships with the full platform.

hobby

$0/mo

for personal projects and prototypes.

100,000 requests / month
7-day log retention
3 workflows
2 team members (you + cofounder)

popular

scale

$49/mo

for teams shipping production apis. from $0.20 per 10k requests above plan.

10M requests / month
30-day log retention
unlimited workflows
unlimited team members

enterprise

custom

for platforms shipping at billions of requests / month.

unlimited requests
custom retention (up to 365d)
sso + scim provisioning
dedicated infrastructure

build a flawless api.
run it like one.

free for the first 100,000 requests. no card. no catch. three lines.