Contract drift monitoring

Catch schema breaks before your agents do.

DriftGuard watches live REST and MCP endpoints, classifies every diff, and sends breaking-only alerts. Start free with the open-source MCP client — add hosted watches when you need continuous checks and a fleet console.

Try it now See live demo
No credit card · 1 endpoint full Pro trial Open-source MCP + CLI Hosted from $39/mo

Monitor the APIs and MCP servers your agents depend on

The watchtower

Catch drift before production does.

Vendor APIs and MCP tool catalogs change on their schedule — not yours. DriftGuard turns silent schema changes into classified, actionable alerts.

Console demo

See it in action.

01 — Open the hosted console with sample watches, live KPIs, and breaking counts — the same views your on-call team uses daily.

Open demo console
Watch fleet

Every dependency, one fleet view.

02 — REST APIs and MCP servers in a single watch list with status dots, sparkline history, and breaking counts at a glance.

Classified diffs

Breaking, warning, and info — not raw JSON noise.

03 — Required field removed pages on-call. Optional param added logs as warning. Description tweaks stay info.

MCP-native

Grounded in your agent stack.

04 — Track tools/list schema changes before Cursor agents break in production. Same watch model as REST OpenAPI.

Alert routing

Wires into the stack you already run.

05 — Slack, PagerDuty, webhooks, and Jira with breaking-only filters. Webhook v2 includes agentAction hints for each change.

Teams we power

Schema monitoring grounded in your workflow.

From agent builders to platform teams — catch external dependency drift before production traffic does.

Vendor APIs.

Stripe, GitHub, Twilio — live OpenAPI watches without manual spec audits.

MCP agents.

tools/list catalog drift before agents break in prod.

CI gates.

Fail builds when production dependencies aren't watched.

On-call.

Breaking-only pages — not description tweaks or HTTP blips alone.

Platform ops.

Portfolio Compass, Insights dashboard, and Needs Review queue in the console.

Platform teams.

200 watches, 5-minute checks, and fleet-wide coverage.
vs CI diff toolsWatches live endpoints you consume, not repos you own.
vs uptime monitorsDetects schema shape changes, not just HTTP 200.
vs manual checksScheduled snapshots with classified breaking/warn/info.
vs raw JSON diffIncident loop, ack/resolve, and alert routing built in.

View all solutions →

Without vs with runtime schema monitoring

Vendor APIs and MCP servers change on their schedule — not yours.

✕ Without DriftGuard

  • Vendor and MCP changes ship silently
  • Issues discovered after users or agents fail
  • Manual OpenAPI diff when someone remembers
  • Pagers on HTTP blips, not schema shape

✓ With DriftGuard

  • Scheduled snapshots with classified diffs
  • Breaking-only alerts to Slack & PagerDuty
  • Incident ack, resolve, and audit trail
  • CI coverage gates and OSS gate packages (MockDrift → SchemaSync)

Open-source gate ladder

Add checks in order — local and CI first, hosted watches when you need always-on monitoring.

Gate 1

MockDrift

Replay drift scenarios in tests before production.

How-to → Gate 2

FuseGuard

Stop runaway agent tool loops and trip on active drift.

How-to → Gate 3

ToolChange

Lint MCP tool manifest changes in PRs.

How-to → Gate 4

SchemaSync

Check prompts against removed schema fields; draft PRs on hosted.

How-to →

Read the gate ladder →

0
Pro watches
0
Team watches
0
Fastest checks
0
Severity levels
0
Audit retention
0
Alert channels
0
Export rows (Pro)
0
History bars / watch

Up and running in minutes

Three steps from signup to your first breaking alert.

1

Add a watch

Paste any HTTPS API or MCP endpoint. Import from catalog or mcp.json.

2

We snapshot & diff

Scheduled checks capture schema shape and compare against the last snapshot.

3
🔔

Alert on breaking

Webhooks fire on breaking drift. Acknowledge and resolve incidents from the console.

Simple, transparent pricing

Self-host free forever, or let us run the watches.

Open source

Self-host

$0
  • compare_json + MCP tools
  • MockDrift, FuseGuard, ToolChange, SchemaSync
  • CLI diff for CI pipelines
Most popular

Pro

$39/mo
  • 50 watches · 30-min checks
  • Webhooks + Slack alerts
  • 90-day history + export
Platform teams

Team

$99/mo
  • 200 watches · 5-min checks
  • PagerDuty + audit export
  • 365-day retention

Resources to get started

Everything you need to evaluate, integrate, and scale monitoring.

Start

Getting started guide

Local-first path: install, diff, MCP, first watch, CI gate.

Getting started → Integrations

Integrations catalog

GitHub Actions, Cursor MCP, Slack, GitLab, and webhooks.

Integrations → CI

Connect GitHub Actions

Trial session, workflow snippet, and coverage assert.

CI setup → Demo

Watch the product tour

Live console with sample watches, drift, and incidents.

Open demo → Hub

Browse all resources

Guides, changelog, activation, and solution playbooks.

Resource hub →

View full resource hub →

Get started with DriftGuard

Unlock hosted watches, MCP monitoring, and breaking-only alerts. Start on one endpoint — scale to fleet-wide coverage when you're ready.

Start building Browse resources