Trust your alertsin production.

Logfox decides what actually deserves your attention, investigates what happened, and helps you fix it.

  • Every component, one place

    APIs, workers, frontends, mobile apps, Lambdas — anything that emits logs or telemetry. Native OTel, CloudWatch, SDK, and more.

  • Setup in minutes

    No YAML to write, no collectors to set up, no learning curve. Connect a service and we walk you through it — minutes, not weeks.

  • Tighten the loop

    Ship with AI at full speed. Logfox watches every environment, flags what matters, and shortens the feedback loop.

Stop firefighting. Start fixing.

Teams waste hours on every incident — context switching, reproducing the problem, and firefighting that eats half the week. Time you meant to spend building. Logfox gives those hours back — detect, investigate, root-cause, and fix in minutes, from the facts that matter, with evidence behind every conclusion.

Click around to explore

Know before your users do

When something breaks, you get a desktop or mobile push — tap it, land in Slack or Discord, and snooze, resolve, or grab a prompt for your own agent without opening another tab.

Open webhook payloads and per-component routing — wire n8n, on-call bots, or your own automations without a middleware SKU.

Real-time logs, zero learning curve

Tail logs across every service as they arrive. Search text and structured fields with forgiving syntax, or filter by environment, service, or level — no query language required.

Bring your own agent

Use Cursor, Claude, ChatGPT, or any MCP client. Logfox is MCP-first: pretty much everything in our API is a tool your agent can call.

Because telemetry is already parsed, normalized, and correlated, agents consume it cheaply — fewer tokens burned, faster answers, better ones.

ChatLogfox MCP

I'm refactoring checkout-web — explain the current issues

I'll check the open alerts and error history for checkout-web.

  • CalledLogfox MCPlist_alerts
  • CalledLogfox MCPget_errors
  • CalledLogfox MCPlist_traces

Three alerts are open, and one dominates:

  • P2 · 3,412 hits — cart.items is undefined when a session resumes from storage. Every trace ends atCheckoutCart.tsx:64.
  • P3 · 210 hits — coupon lookup times out at 5s, masked by retries.
  • P3 · 44 hits — stale price shown when a promo expires mid-session.

The P2 is 92% of the volume. It started 9 days ago with v2.11.3, which changed how the cart hydrates — guard that path first.

Plan, search, build anything
Agentclaude-4.5-sonnet
Coming soon

Send a metric, it shows up

Send a metric from the SDK or API — a name and a number. Logfox charts it and puts it on your dashboard by itself — nothing to declare first, no YAML, no PromQL.

Rearrange the ones you have, switch between line and bars, and set an alert that opens pre-filled with a sensible threshold. Same SDK and API for alerts when you want them in code.

Coming soon

See what your architecture actually looks like

Discovered automatically from the telemetry you already send. See what talks to what, spot the places where errors pile up, and jump straight into its logs.

Service map

ProductionSearch services/
Auto-discovered · 7 services · 6 edges · 1 vendor
Coming soon

Distributed tracing: follow any request end to end

One flame graph per request: every service it touched, in the order it touched them. Click a span to read the logs it emitted — inside your logs view, under the filters you already set. No second tool, no query language.

Logs

Trace 4f2b9c1e · 320ms · 6 spans · 3 errors
Search logs/
Live tailTrace
ProductionAll servicesInfo+
POST /checkoutedge · 320ms2 of 2 logs match your filters
02:30:41.002edgeINFOcheckout started cart=cart_8812 items=3 request_id=req_7fd2
02:30:41.322edgeERROPOST /checkout status=500 latency_ms=320 request_id=req_7fd2
POST /chargepayments-api · 269ms3 of 5 logs match your filters
02:30:41.020payments-apiINFOcharge requested amount=4200 currency=usd customer=cus_9f3
02:30:41.184payments-apiWARNretry scheduled attempt=2 route=/v1/charge delay_ms=800
02:30:41.288payments-apiERROStripeError: upstream returned status 500 at charge (src/stripe/client.ts:142)
GET /token/verifyauth-api · 38ms1 of 3 logs match your filters
02:30:41.045auth-apiINFOJWT verified user=usr_8f2 scopes=charge.write
SET idempotencyredis · 13ms0 of 2 logs match your filters

No logs match your filters.

INSERT chargespostgres · 51ms1 of 1 logs match your filters
02:30:41.108postgresINFOINSERT INTO charges rows=1 duration_ms=48
POST /v1/chargesstripe · 141ms2 of 4 logs match your filters
02:30:41.152stripeINFOPOST /v1/charges idempotency_key=idem_9c4
02:30:41.279stripeERROcard_declined decline_code=insufficient_funds status=500

  • Sensitive data, redacted at ingest

    PII and secrets are redacted by meaning at ingest — strong normalization plus layered regex — so sensitive values never touch disk and never reach a model.

  • Integrations & collectors

    Keep your existing collectors or use ours: OpenTelemetry, CloudWatch, the Logfox SDK, or plain HTTP. Or skip all that — logfox run wraps any command and streams its logs in.

  • 60-day retention for what matters

    Raw telemetry rolls off after 7 days. Important samples, rollups, and investigations remain available until 60 days after the alert is resolved.

Affordable pricing for teams of all sizes

One simple bill. AI included. No surprises.

Save 17%

  • Free

    $0/mo

    • Seats1
    • Components3
    • Telemetry10 GB
    • AI summaries50
    • AI investigations5

  • Starter

    $99/mo

    • Seats5
    • Components15
    • Telemetry100 GB
    • AI summaries1,000
    • AI investigations50
  • Popular

    Growth

    $299/mo

    • Seats15
    • ComponentsUnlimited
    • Telemetry500 GB
    • AI summaries10,000
    • AI investigations200

  • Business

    $899/mo

    • Seats75
    • ComponentsUnlimited
    • Telemetry2 TB
    • AI summaries50,000
    • AI investigations500

  • Dedicated

    From $999/mo

    An isolated, single-tenant Logfox server, operated by us. Seats and telemetry sized to your workload. SSO included. Bring your own model — you provide the API key. Billed annually.

    Contact us

  • Self hosted

    Custom

    Run Logfox on your own infrastructure — your data never leaves your network. Custom retention and security available. SSO included. Bring your own model — you provide the API key.

    Contact us

Tiers may shift before launch. Slack and Discord alerts don’t consume AI quotas.

Know what matters.Ignore the rest.

Ship faster and sleep better. Logfox watches every environment, wakes you only when something real breaks, and hands you the cause and the fix.

No spam. Unsubscribe anytime.