Apperio watches every error, every deploy and every commit in between, then tells you in plain English which change caused the break and what to do about it. One npm install. Nothing to configure.
You wrote it, you tested it, you shipped it. Then it met real browsers, flaky networks, and users doing things you never imagined. From that point on you are guessing, and your only monitor is a customer annoyed enough to write in.
What changed
a7f3c21 fix: cache user profile
3b91e04 chore: bump deps
c2d7a18 feat: express checkout
What broke
TypeError /checkout
Slow GET /api/users 1.8s
Failed POST /api/pay
Both halves of the answer already exist. Nothing in your stack is holding them in the same hand.
By the time someone bothers to email you, the error has been firing for hours. You have no idea how many people hit it first and quietly closed the tab.
Cannot read properties of undefined at chunk-4f2a.js:1:88213. Minified, contextless, and several commits removed from whatever actually caused it.
Your error tracker has never seen your commits. Your git log has never seen production. You reconcile the two by hand, from memory, at 2am.
Enterprise observability is priced and designed for an on-call rotation, a platform engineer, and a procurement cycle. You are one person shipping on a Sunday.
Not another dashboard replaying what already happened. Apperio joins the two halves your stack keeps apart, then does the reading for you.
Apperio ingests your pushes, deploys and releases the second they happen. When an error group appears, it ranks the commits most likely to have caused it by how far their changed files overlap the failing stack, then shows the winner as Likely caused by, with the diff and the deploy it rode in on.
a7f3c21 fix: use cached user profile on checkout
92%Touched checkout.ts, the top frame in the stack
3b91e04 chore: bump dependencies
21%No overlap with the failing path
Every commit gets a summary a non-engineer can follow. Every error group gets a root cause in the same voice. Every deploy gets a verdict: better, worse, or no change. Flip to the technical register when you want frames and line numbers instead.
Go from an error group straight into the recording of the exact session that produced it, with the clicks, scrolls and route changes that led there. Every text input and password is masked before it leaves the browser. Off until you switch it on, per project, seven-day retention.
Session started
This is not a highlight reel of separate features. It is one Tuesday afternoon, in order, with nothing configured in advance.
The webhook reaches Apperio before your terminal has finished printing. The commit is stored, the diff is read, and a summary is written in the register a non-engineer can follow.
fix: use cached user profile on checkout
3 files changed +41 -12
Summary: speeds up checkout by reusing the profile the page already loaded instead of fetching it a second time.
Vercel, GitHub Deployments, your CI, or one API call from a shell script. However the release happens, Apperio draws the line on your error, performance and activity charts so every number after it is attributable.
v2.4.1 -> production (github)
Baseline captured. Watching for 60 minutes.
The SDK catches the throw with the breadcrumbs that led to it: the route, the clicks, the network calls, the release tag, the session. No try/catch of yours involved.
TypeError: Cannot read properties of undefined (reading 'email')
at renderCheckoutSummary (checkout.ts:142)
release v2.4.1 · session 8f2k4n · Chrome 122
Not a wall of 14 identical rows. One group, deduplicated by the shape of the failure, with how many people it has reached, when it first appeared, and which release it appeared on.
14 users affected · first seen 2 min ago
100% on v2.4.1 · 0% on v2.4.0
In-app and by email. Nobody configures a threshold for an error that has never existed before. A new kind of failure is worth interrupting you for, and that is the default.
The failing stack points at checkout.ts. One commit in this release touched checkout.ts. Apperio ranks the candidates, puts that one at the top, and shows you the exact lines it changed.
a7f3c21 fix: use cached user profile · 92%
- const profile = await fetchProfile(userId)
+ const profile = cache.get(userId)
Title, plain-English summary, stack trace, breadcrumbs, affected users, suspect commit and a link back. Edit anything you like in the preview, then open it on your repo.
#142 Checkout fails for accounts with no email set
opened by apperio[bot] · labels: bug
The status flows back on the webhook. The group goes quiet, the release health verdict updates, and if the same fingerprint ever returns, it reopens itself and tells you.
v2.4.2 deployed · error rate -100% vs v2.4.1
Verdict: this deploy made things better.
The only thing you did in those thirty-one minutes was read, click once, and ship the fix. Everything above it happened because the SDK was installed and the repo was connected.
No agent to run, no collector to host, no YAML. If you can install a package and click Authorize, you are done.
npm install apperioimport { Apperio } from 'apperio'
new Apperio({
apiKey: process.env.APPERIO_KEY,
environment: 'production'
})That is the whole integration. Errors, network calls, clicks, page views, console output and Web Vitals start flowing immediately, with PII stripped before anything leaves the browser.
you/shop
Apperio App installed
push · deployment · release · issues
Backfilled 50 commits
Install the GitHub App on the repos behind the project. Commits, deploys and releases stream in over webhooks from that second, and your recent history is backfilled so the feed is never empty.
Optional. Apperio is a complete error and performance monitor without it. The repo is what adds the change side of the story.
This week on you/shop
Open it when something needs you. Otherwise let the digest come to you: what you shipped, what it did to your users, and the one thing worth looking at next.
One SDK, one dashboard, one source of truth. Nothing here is an add-on tier or a second product.
Logs are noise without context
Anomaly surfacing, root cause explanation and fix suggestions, written against your actual stack trace rather than a generic pattern library.
Copy-pasting stack traces into issues
One click turns an error group into a GitHub issue with everything attached. Close it there and Apperio marks the error resolved here.
Checkout fails for accounts with no email set
Stack, breadcrumbs, 14 affected users and the suspect commit, attached automatically. Close it here and the error resolves in Apperio.
Users leave before they complain
Web Vitals, network timing and interaction latency, captured on every real session rather than a synthetic run from a data centre.
Manual logging is always incomplete
Errors, network requests, clicks, page views, console output and Web Vitals, captured from the moment you initialise. No wrapping, no decorators.
Accidentally logging your users
Ten-plus detection patterns for emails, cards, national ID numbers, API keys and JWTs, redacted in the browser with an audit trail of every substitution.
Finding out on social media
Slack, Discord, Teams, Linear, Jira, PagerDuty, email or a raw webhook. Thresholds and anomalies when you want them, sensible defaults when you do not.
Error spike in /api/checkout — 23 in 5 min, baseline 2
Did that deploy help or hurt?
Every deploy gets a baseline and a verdict. Better, worse, or no change, measured against the release before it rather than left to your memory.
Nobody opens the dashboard
Weekly or daily, in plain English: what you shipped, what it did to your users, and the one thing worth looking at next.
You shipped 11 commits across 3 deploys. Errors fell 47%. Your slowest page is /settings at 2.4s, and it got worse on Thursday.
Apperio does not assume you have a platform team. It assumes you have something in production and not enough hours in the day.
You are the entire on-call rotation. Apperio tells you what broke and which of your own commits did it, so your evening is a one-line fix and not a three-hour bisect.
Zero-config error notifications
You shipped code you did not write line by line. A plain-English summary of every commit means you always know what actually changed in your app, whoever or whatever typed it.
Every commit, explained
Errors grouped by cause, the right owner notified automatically, and the issue drafted in your repo with full context. Triage stops being a meeting.
Issues drafted, status synced
A project per client, an API key each, one account. Send the weekly digest instead of writing the status update, and answer what broke before the client asks.
Per-project isolation
Release health on every deploy, so you know within minutes whether the thing you just shipped made it better or worse, with the baseline measured for you.
Deploy impact verdicts
An abandoned session with an error attached is revenue walking out. Replay the exact session, find the frame it died on, ship the fix before the day ends.
Session replay from an error
Everything in this table can tell you what broke. Apperio is built to answer the next two: what did we ship that broke it, and what do I do now.
| Capability | Apperio | Error trackers | APM suites | Platform logs |
|---|---|---|---|---|
| Errors grouped and deduplicated | ||||
| Commits and deploys on the same timeline | Built in | Release tags | Deploy markers | |
| Ranked suspect commit for each error | AI-ranked | |||
| Plain-English explanations | Commits and errors | Error summaries | Incident summaries | |
| Verdict on whether a deploy helped | Automatic | Manual comparison | ||
| Issue drafted with full context | One click, synced | Link only | ||
| PII stripped before it leaves the browser | 10+ patterns, audited | Server-side scrubbing | ||
| Usable by a non-engineer on your team | ||||
| Time from install to first insight | Under 5 minutes | An afternoon | Days | Already on |
| Affordable for one person | Free tier, forever | Free tier with caps | Team pricing |
Compared by category rather than by brand, because the answer varies within each one. If your current tool does something on this list, keep it. Apperio installs alongside anything.
Most of what this page describes is already running. Here is the honest state of the build, kept current as things land. No screenshots of features that do not exist.
Installation tokens, signed webhook receiver with delivery-ID idempotency, and stored commits, deploys and releases per project.
Plain English and technical, cached per SHA, with lockfiles and generated files filtered out of the diff.
Deployments from GitHub, your CI, or a single API call, drawn onto the charts with an automatic better-or-worse verdict.
Errors grouped at ingestion, counted by users affected, and routed to the owner in-app and by email without an alert rule.
Editable preview, full context attached, opened on GitHub in one click. Close it there and the error group resolves here.
Stack-file overlap plus AI ranking, computed on demand and surfaced as Likely caused by on the error group.
Recorder in the SDK, player on the session page, all text inputs masked by default, opt-in per project, seven-day retention.
One chronological feed per project, and a weekly or daily email that reads like a person wrote it.
First-run flow that gets you from signup to first captured error without reading documentation.
The beta is small on purpose. Every account added is one more app whose real traffic shapes what ships next.
The first accounts through the door are charter members. This is not a launch discount that expires. It is a standing arrangement for the people who show up before it is obvious.
Whatever tier you land on during the beta stays at that price for as long as you keep the account. Public pricing will be higher than this.
Every feature unlocked while the beta runs, including the AI layer, Change Intelligence and session replay, on every project you create.
A private channel to the person building this. Bug reports skip the queue, and you get an answer rather than a ticket number.
You see the roadmap before it is public and you vote on what gets built next. Charter feedback outranks the backlog.
Not a trial that lapses into a paywall. A real free tier with a monthly log allowance, live projects and alerts, meant for the side project that may never make money. Paid tiers start at nine dollars a month when you outgrow it.
Invites go out in batches, in waitlist order, as capacity allows. No card, no call, no demo to sit through.
The SDK is TypeScript and runs in browsers and in Node today, with no HTTP client dependency. Underneath it is a plain HTTPS endpoint, so anything that can POST JSON can send logs to a project right now. First-party SDKs for other languages come after launch.
No. Apperio is a complete error, performance and session monitor without it. Connecting a repo is what adds the change side of the story: commit summaries, deploy markers, suspect commits and drafted issues. You can add it later, and your recent history gets backfilled when you do.
The SDK redacts personal data before anything leaves the browser: emails, credit cards, national insurance and social security numbers, API keys, JWTs and more, across ten-plus built-in patterns, with an audit trail of what was redacted and why. You can add your own rules or pick a stricter preset. Session replay masks every text input and password by default and stays off until you turn it on per project.
There is a free tier and it is not a trial. Paid tiers start at nine dollars a month and scale on log volume rather than seats, so adding a teammate never costs more. Charter members keep whatever price they joined on.
Do not switch yet. Apperio installs in one line and does not conflict with anything else on the page, so run both for a fortnight on the same app and compare what each one told you the next time something broke. If the answer is the same, keep what you have.
The SDK batches log submissions rather than sending one request per event, with a configurable batch size and flush interval, exponential backoff with jitter on retry, and a bounded in-memory buffer so a network outage cannot grow without limit. It ships as both CommonJS and ESM with no runtime HTTP dependency.
Commit diffs for summarisation, with lockfiles and generated files filtered out and large diffs truncated, plus the error group being explained. Summaries are cached per commit SHA so the same diff is never sent twice, and there is a spend cap per project. If the AI layer is unavailable, everything else keeps working and the summary simply says so.
Invites go out in batches in waitlist order as capacity allows. You get an email with a code that unlocks signup. Sharing your invite link moves you up the queue when someone joins through it.
Something else on your mind? femi@apperio.dev reaches a person, not a form.
Every hour an error runs unseen is a person who tried your product once and decided it was broken. They will not email you. They will just not come back.
npm install apperioFree tier forever · No credit card · Charter pricing locked · One-click unsubscribe