Developers

Xovris in your code and pipeline

Finds what broke, fixes it on a copy, opens the pull request.

Start freeRead the API reference

Send a requestapi.xovris.ai

GET /v1/public/catalogue/totals answered 200 in 767 ms

{
  "services": 14533,
  "byState": {
    "verified": 0,
    "supported": 1662,
    "generated": 12360,
    "browser": 0,

+34 lines – send it above for today’s

+30 lines – send it above for today’s

api.xovris.ai’s own reply, read 9 October 2026 at 21:21 UTC

Libraries

MCP server

https://api.xovris.ai/v1/mcp Add it to your assistant, client by client

The same Xovris, from your own code

Start a check, follow an investigation, read the proof and stop the next problem before it starts, from a script or a deploy pipeline. Every request on this page is one the API answers today; send a public one right here.

Read the API reference

Check, investigate, fix, prove and prevent – from your own code

Everything the command centre does, the API does: start a check, follow an investigation, approve the exact fix you were shown, read the proof that it worked, and let Xovris stop the next problem before it starts. Each request below is one the API answers today.

Check

Add a check in plain words, check for exposure, or have a query checked before you run it.

  • POST/v1/checksAdd a check
  • POST/v1/exposure/checkCheck for exposure now
  • POST/v1/query/checkExplain a query and check it for mistakes
  • GET/v1/checksList checks
curl -X POST https://api.xovris.ai/v1/checks \
  -H "Authorization: Bearer $XOVRIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"describe":"check https://shop.yourcompany.com every 5 minutes"}'

Investigate

Read every incident, every cause Xovris is weighing and what tested each, and add what you know so the investigation uses it.

  • GET/v1/incidentsList incidents
  • GET/v1/incidents/{id}/planRead every cause Xovris is weighing for an incident and what tested each
  • GET/v1/incidents/{id}/dependenciesRead what an incident's asset rests on, how recently each link was seen, and how that lowers the diagnosis
  • POST/v1/incidents/{id}/notesAdd a note to an incident (a paused investigation wakes on it)
  • GET/v1/issuesList errors
curl -X GET https://api.xovris.ai/v1/incidents \
  -H "Authorization: Bearer $XOVRIS_API_KEY"

Fix

Ask for a fix, then approve the exact version you were shown – every change follows your own approval settings.

  • POST/v1/incidents/{id}/repairsAsk Xovris to fix an incident
  • POST/v1/repairs/{id}/approveApprove a fix
  • POST/v1/incidents/{id}/choiceChoose what happens next on an incident
  • POST/v1/vulnerabilities/upgrades/{id}/fixAsk Xovris to make an upgrade
curl -X POST https://api.xovris.ai/v1/incidents/ID/repairs \
  -H "Authorization: Bearer $XOVRIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

Prove

Read each fix with its before and after, the record of everything Xovris did, and a check that the record was never altered.

  • GET/v1/actionsList what Xovris did and why
  • GET/v1/actions/verifyVerify the record of what Xovris did has not been altered
  • GET/v1/incidents/{id}Read an incident
curl -X GET https://api.xovris.ai/v1/actions \
  -H "Authorization: Bearer $XOVRIS_API_KEY"

Prevent

Read what will fail before it does, choose what Xovris fixes on its own, and watch a domain or brand for exposure.

  • GET/v1/forecastsRead what will fail before it does
  • PATCH/v1/settings/autonomy/{kind}Change what Xovris may fix on its own, for one kind of change
  • POST/v1/exposure/watchWatch an address, domain or brand
  • GET/v1/vulnerabilitiesList software vulnerabilities that affect you
curl -X GET https://api.xovris.ai/v1/forecasts \
  -H "Authorization: Bearer $XOVRIS_API_KEY"
  • API tokens A scoped, expiring API token (Settings, Developers) or a session token.
  • Signed-in sessions A signed-in person's session.
  • locationToken A check location's own token, issued when the location is enrolled.
  • extensionPairing The browser extension's pairing token, issued by POST /v1/doors/extension/pair.
  • webhookSignature The code host's signature of the delivery (each host's own header), checked against the secret set when the hook was made.
  • connectorToken The connector's own token, shown once when the connector is added.
  • internalToken Xovris's own services only.
  • deskKey A cloud desktop's own key, given to it when it starts - its two doors only.
  • ingestKey The project's DSN key - a Sentry SDK sends it in X-Sentry-Auth (sentry_key), or as sentry_key in the query.
  • datadogApiKey The workspace's Xovris ingest key, carried in the Datadog Agent's DD-API-KEY header.
  • partnerProvisioningKey A partner's provisioning key (Partners, Provisioning keys) - it signs this one call alone.

Each example reads your key from XOVRIS_API_KEY, so the key never sits in your terminal history. Replace ID with the record you mean.

In your code and your pipeline

Every way in from your code and pipeline that answers today, each with its exact start. The rest is on What’s next for Xovris.

Learn more
  • The setup wizard

    Finds your framework, signs you in through your browser, shows every change before it writes one and checks the first event arrives.

    npx xovris@latest init
  • SDKs for Node.js, Next.js and the browser

    Errors, unhandled rejections and traces, joining an OpenTelemetry setup you already have instead of starting a second one.

  • Client libraries

    Built to match the published API description: TypeScript, Python and Go.

Webhooks you can trust

A free kit that checks your server is ready to work with Xovris. It needs Node.js 18 or later and nothing else, and it is yours to keep under the MIT licence.

Your webhook receiver, checked by Xovris before you go live

Run it with your endpoint’s signing secret in the environment, so it never lands in your shell history:

curl -O https://xovris.ai/developers/xovris-webhook-conformance.mjs
XOVRIS_WEBHOOK_SECRET=whsec_... node xovris-webhook-conformance.mjs https://hooks.yourcompany.com/webhooks/xovris

Download the kit

It sends seven requests and passes your server only when it:

  1. accepts a correctly signed event;
  2. acknowledges the same event sent again, so retries stop;
  3. refuses an event whose body was changed after signing;
  4. refuses an event with no signature;
  5. refuses a signed event that is 10 minutes old, so a replay fails;
  6. refuses an event signed with another secret;
  7. answers within 10 seconds – a 202 is fine.

Keys that are safe to keep

Create a key for each script, pipeline or tool, so you always know what is calling and can stop it on its own.

Keys you control

  • Read, or read and change – each key holds exactly the access it was given.
  • You choose how long each key lasts, from a day to a year.
  • The key is shown once. Xovris keeps only a fingerprint of it, never the key itself.
  • Revoke a key and the very next call with it is refused.
  • Rotate a key with one press: a new key with the same choices is shown once, and the old one is revoked.
  • Before a key exists, you read in one paragraph what it will be able to do, where, for how long and from where.
  • Keys limited to the actions you choose
  • Keys for one domain, site or server
  • Keys that never end
  • Keys usable only from your addresses
  • A spending limit for each key
  • Test keys: a test key runs every request against a copy, so the real things stay as they are while you build.
  • Every call logged against its key
  • An alert when a key is used unusually
  • A change made through the API follows your own setting for that kind of change, exactly as one made in the command centre: Xovris makes it on its own, shows it to you first, or leaves it alone – you choose.
  • Every kind of change names how it is undone before it runs; a change with no way back is refused, never made.

Safe, without holding you back

  • A change asked for through the API follows the same choices you made in the command centre – on its own, review first, or off.
  • Every change carries its way back, and Xovris puts things back on its own if the change does not work.
  • A key holds exactly the access it was given, whatever it asks.

One price for each outcome, wherever it starts

A check, an investigation or a fix started through the API is the same one you would start in the command centre – counted against the same plan, at the same price, and with no charge for a fix that does not hold.

  • A check, an investigation or a fix through the API counts against the same plan at the same price as in the command centre – the command centre itself runs on these requests.
  • You pay for a fix that works, never for an attempt: a fix that does not hold costs nothing.
See the prices

The API runs the same AI as the command centre and is held to the same standard.

Everything the API covers

Every area of the command centre is an area of the API. Each is described in the reference with every request it takes, the token it needs and what it answers.

Sending data
Send errors, logs, metrics and traces from your apps, or from the tools you already run – OpenTelemetry, Sentry SDKs, and Datadog and Elastic log shippers – with your project’s ingest key.
Incidents
Read each incident with its story, evidence and timeline, and choose what happens next.
Fixes
Approve, decline or ask for a change to a fix Xovris has prepared. A fix runs only on the exact version you approved.
What Xovris did
Every read, change, decision and rollback Xovris recorded, with its reason and its way back; nobody can alter the record unnoticed.
Asset stories
The whole life of a certificate, a domain or a server, and the same story as a file anyone can verify offline.
Errors
Your apps’ errors, grouped, with their counts.
Sentry issues
The unresolved issues of a Sentry organisation you connected, each linking back into Sentry.
Checks
Add checks of websites, APIs and journeys in words or fields, choose how often they run, pause them and read every result.
Sites
Add a website by its address and read its first look and its quality report.
Free checks
The free checks on xovris.ai, callable from your own scripts within their limits.
Logs
Search your logs in plain words or fields, open a line in its context, and read patterns and retention.
Log pipelines
Rules that shape your logs on the way in, and the series and monitors derived from them.
Log archive
Where older logs are kept, and searching them.
Forwarders
Where your logs are also sent, and whether every copy arrived.
Metrics
Your systems’ measurements, what Xovris noticed in them, and queries over them.
Forecasts
What is likely to fail before it does – a computer’s disk or battery, a domain or a certificate running out – each with the fix scheduled ahead of it.
Dashboards
Dashboards you made or Xovris made for you, their panels and their silences.
Devices
The computers running the Xovris app, their health and what happened on them.
Software vulnerabilities
Vulnerable software on your computers, the one upgrade that closes each, and asking Xovris to run it.
Breach and lookalike watch
Your addresses, domains and brand watched for breaches and imitations, and the response to each.
Cyber Essentials readiness
Each requirement as met, not met or not read yet, from your real computers and accounts.
Connections
The services you connected, what each lets Xovris read and change, and their credentials’ life.
Tools you connect
Hand incidents to the tools your team uses, and read every delivery.
Payments (Stripe)
Failed payments and disputes read as a story, with refunds and replies that wait for you.
Deployments
Record your deployments and read how often and how safely you ship.
Automation
What Xovris may change on its own: the access you granted, your standing permissions, stop switches, change windows and the approvals waiting for you.
Settings
Your workspace’s AI limit and how much Xovris may do on its own, per kind of change.
Setup
Where setup stands, each step’s state and its health.
Webhooks
Endpoints Xovris sends signed events to, every delivery and its retries.
API tokens
Create, list and revoke the scoped, expiring tokens these calls use.
Members
The people in your workspace.
Workspaces
Create a workspace.
You
Who the token or session belongs to.
Export your data
Everything Xovris holds for your workspace, in one download.

And 151 more areas – 937 calls in all – each in the reference with every request it takes:

  • Alerts 61 calls
  • Code 57 calls
  • Autonomy 56 calls
  • Reviews 42 calls
  • Assistant 36 calls
  • Connected systems 28 calls
  • Computer 25 calls
  • App store watch 24 calls
  • On call 24 calls
  • Identity systems 20 calls
  • Computers 19 calls
  • Intrusion detection 19 calls
  • Network 18 calls
  • Web filtering 18 calls
  • Sessions 17 calls
  • OAuth apps 15 calls
  • Partners 14 calls
  • SCIM 14 calls
  • Real-user monitoring 12 calls
  • Workspace 12 calls
  • People and access 11 calls
  • Cloud 10 calls
  • Status pages 10 calls
  • Things 10 calls
  • Traces 10 calls
  • Doors 9 calls
  • Inbound webhooks 9 calls
  • Integrations 9 calls
  • Log connections 9 calls
  • Extensions 7 calls
  • Security stack 7 calls
  • Single sign-on (SSO) 7 calls
  • Watches 7 calls
  • Backups 6 calls
  • Board 6 calls
  • Collectors 6 calls
  • Mail 6 calls
  • MSP 6 calls
  • Procedures 6 calls
  • Protocol connections 6 calls
  • Roles 6 calls
  • Security 6 calls
  • Status page 6 calls
  • Storage 6 calls
  • Teams 6 calls
  • Uploads 6 calls
  • Connect 5 calls
  • Device sign-in 5 calls
  • Goals 5 calls
  • Queries 5 calls
  • Account 4 calls
  • Alert actions 4 calls
  • Carrying on 4 calls
  • DNS 4 calls
  • Elevations 4 calls
  • Everyday 4 calls
  • People 4 calls
  • Profiles 4 calls
  • Telemetry 4 calls
  • Tool servers 4 calls
  • Agents 3 calls
  • Badges 3 calls
  • Efficiency 3 calls
  • Exports 3 calls
  • Failover targets 3 calls
  • Free 3 calls
  • Grants 3 calls
  • Helper 3 calls
  • Home network 3 calls
  • Invitations 3 calls
  • Make 3 calls
  • MCP 3 calls
  • Models 3 calls
  • Outcomes 3 calls
  • Payments 3 calls
  • Projects 3 calls
  • Role grants 3 calls
  • Scam check 3 calls
  • Searches 3 calls
  • Vendor fixes 3 calls
  • Views 3 calls
  • What if 3 calls
  • Builds 2 calls
  • Changes 2 calls
  • Companies 2 calls
  • Decoys 2 calls
  • Developer sandbox 2 calls
  • Disposal 2 calls
  • DNSSEC 2 calls
  • Edge 2 calls
  • Everything you run 2 calls
  • Incident scopes 2 calls
  • Insider 2 calls
  • Live 2 calls
  • Major incident updates 2 calls
  • Query 2 calls
  • Readings 2 calls
  • Recoveries 2 calls
  • Travel 2 calls
  • Watch 2 calls
  • Agent to agent (A2A) 1 call
  • AI 1 call
  • API change watch 1 call
  • API reference 1 call
  • Audit 1 call
  • Billing 1 call
  • Bills 1 call
  • Breach notices 1 call
  • Budget 1 call
  • Capacity 1 call
  • Choices 1 call
  • Continuity 1 call
  • Conversations 1 call
  • Crypto 1 call
  • Data 1 call
  • Decisions 1 call
  • Deploy guard 1 call
  • Digest 1 call
  • Domains 1 call
  • Fix actions 1 call
  • Footprint 1 call
  • Health 1 call
  • History 1 call
  • Ingestion keys 1 call
  • Laws 1 call
  • Needs you 1 call
  • Now 1 call
  • OpAMP 1 call
  • Passwords 1 call
  • Permissions 1 call
  • Phishing 1 call
  • Pins 1 call
  • Post-mortems 1 call
  • Prometheus 1 call
  • Protocol changes 1 call
  • Provisioning 1 call
  • QR codes 1 call
  • Reports 1 call
  • Rescue kit 1 call
  • Secret scanning 1 call
  • Shell 1 call
  • Sign-in 1 call
  • SIM swap 1 call
  • SLA credits 1 call
  • Spend 1 call
  • Statement 1 call
  • Status sentence 1 call
  • Subscriptions 1 call
  • Suppliers 1 call
  • Tools 1 call
  • Value 1 call

Build on Xovris

  • Make an internal integration in one press: its key and its signed webhook together, each secret shown once, live at once with no review.
  • A slow server is never punished: Xovris waits 15 seconds for an answer and counts a 202 as delivered, then retries a failure on a published schedule with the same webhook-id.
  • Put your own tool inside the command centre in eight ways – an issue link, an alert action, an asset panel, a fix review, a settings page, a timeline entry, a status badge or a sandboxed app frame – each checked against a published schema before it is drawn.
  • Send alerts from Sentry, Datadog, Prometheus Alertmanager, Grafana, PagerDuty or any HTTP call, each checked for its signature and turned into one investigation.
  • See how each connected integration is doing from what it delivered – marked degraded at 5 failed deliveries in 100, never hidden.
  • Send any delivery again from the dashboard or the API
  • Publish an app for every Xovris customer
  • Alerts that call your product
  • A safe retry on every change
  • A sandbox with a made-up estate

Ask it anything you run. It sets itself up, asking first.

It answers in plain words: what broke, where, what it changed and what it checked.

Ask Xovris about anything you run

TerminalNext.js

npx xovris init
Xovris – connect your appWhat Xovris foundApp: Next.js - A real install with npm from the public registry passed on 2026-11-15, but not yet on the newest next release (16.3.8).Supported: next 16.0.0 is in the supported range (from 15.3, below 17).Supported: Node.js 22.22.0 on this machine is in the supported range (20.19 or newer).Found: npm (from package-lock.json)Xovris will: Install Xovris into this Next.js app with npm, write its settings and send a test event. Next.js is supported here and every version check passed.Also possible: Watch the app from the outside by its web address. Works for any app on any stack, with nothing installed.Found Next.js, using npm.Waiting for you to approve in the browserSign inYour browser opened to approve this sign-in.Check the browser shows the same code: VJHR-KTCSYour data is never used to train our AI.

npx xovris init

Recorded from npx xovris init on the Next.js test project in the install suite, 15 November 2026, exit code 0

Build on Xovris today

Start free, connect what you run, and Xovris starts watching it.

Start freeStart from the API reference

Ask Xovris

Open the full page