SELF-HOSTED AI-POWERED LOG FILTERING

Reduce Alert Noise.Cut Unnecessary Incident Tickets.

LogsFast adds an intelligent filtering layer between your monitoring tools and your alerting systems. It masks detected sensitive data locally, scores incoming events from 1 to 5 using AI, and filters out below-threshold noise before escalation.

Early Access Offer: Free for the first 10 teams for 30 days in exchange for feedback. Just sign up — no credit card required.

Start 14-Day Free Trial

Keep your existing monitoring and log storage. Let LogsFast help your engineering team spend less time handling unnecessary alerts and more time resolving meaningful incidents.

MONITORING & ALERT SOURCES
Splunk Datadog Grafana CloudWatch Custom HTTP sources
LOGSFAST
Local masking → AI severity scoring → Threshold filtering
Below threshold: Dropped by LogsFast At or above threshold: Forwarded to your webhook

Your destination tool handles ticket creation, notifications, and escalation.

ANY DESTINATION OUT
ServiceNow Jira PagerDuty Slack Teams Opsgenie Webhooks

THE PROBLEM

Alert Noise Is Creating Unnecessary Work for Engineering Teams.

Your monitoring tools detect problems. But not every log containing an error or timeout represents an incident that needs an engineer's attention.

The Alert Overload

High volumes of warnings, retries, and recurring errors can create unnecessary notifications that compete with genuinely important incidents.

Unnecessary Incident Tickets

Every noisy event that reaches your ticketing workflow can create additional tickets, increase triage work, and add operational overhead.

Keyword Rules Lack Context

A harmless background retry and a genuine service failure can contain the same word, such as "timeout." Keyword matching alone may not distinguish their severity.

Manual Triage Takes Engineering Time

Engineers and NOC teams spend valuable time reviewing repetitive alerts instead of focusing on service reliability and meaningful incidents.

No Visibility into Alert Priority

Without severity classification before escalation, low-priority noise and critical production failures enter the same alerting pipeline with equal urgency.

THE OPERATIONAL DILEMMA

Standard Alerting vs. Context-Aware Log Filtering

Why forwarding every alert directly to incident-management tools creates unnecessary triage work—and how pre-escalation filtering helps.

THE PROBLEM

The Log Noise Problem

Traditional keyword-based rules can trigger alerts for routine retries, temporary errors, and harmless background events. These alerts can create unnecessary tickets and make important incidents harder to prioritize.

THE BUSINESS IMPACT
  • Alert Fatigue: Repeated low-value notifications demand attention throughout the day and night.
  • Ticket Overload: Unnecessary events create additional incident records and triage work.
  • Lost Engineering Focus: Skilled engineers spend time reviewing noise instead of resolving important problems.

The Outcome: High Cost, Low Signal, Severe Alert Blindness.

THE LOGSFAST SOLUTION

Filter by Context. Escalate by Severity.

LogsFast sits between your monitoring tools and your destination webhook. It masks detected sensitive values locally, sends the masked text for AI severity scoring, and applies your configured threshold before forwarding the original log and its score.

Incoming Alerts: HTTP POST from your monitoring tools
▼
Step 1 — Local Data Masking (Detects & masks supported sensitive values)
▼
Step 2 — AI Severity Scoring (Masked text evaluated by Typesafe Jev · Score 1–5)
▼
Step 3 — Configurable Threshold (Compare score against alert level)
Below threshold
Not forwarded · Dropped without storage by LogsFast
At or above threshold
Forwarded · Original log + severity score sent to webhook

If AI scoring is unavailable, the default fail-open behavior forwards the log without a score.

Local Sensitive-Data Masking: Detected sensitive values are masked before log text is sent to the external AI scoring service.

Reduce Downstream Noise: Below-threshold events are dropped before reaching your ticketing or alerting workflow.

Keep Your Existing Stack: Connect through HTTP POST and a destination webhook. Keep your current monitoring and log storage systems.

WHAT LOGSFAST DOES

Built to Filter. Designed to Fit Your Infrastructure.

LogsFast combines local processing, AI-assisted classification, and configurable routing in one self-hosted Docker container.

1. Local Sensitive-Data Masking

Detects and masks supported credentials, tokens, email addresses, IP addresses, and other sensitive patterns before AI evaluation. Original logs are retained for forwarding to your own webhook when they qualify.

2. AI-Powered Severity Scoring

Uses Typesafe Jev with your own API key to assign each incoming event a severity score from 1 to 5, from harmless information to critical system issues.

3. Configurable Alert Thresholds

Choose an alert level from 1 to 5. Events below your chosen level are dropped, while events at or above it are forwarded to your configured webhook. Your filtering policy, your choice — pass every log through or filter by your preferred severity threshold.

4. Custom Rules & Webhook Routing

Add plain-English classification guidance for your services and route qualifying events to your chosen HTTP webhook. Your receiving tool controls its own ticketing and notification workflows.

5. Built-In Health & Reliability Controls

Monitor application health, request limits, memory warnings, and AI availability. Docker restart policies and configurable fail-open behavior help manage operational interruptions.

Self-hosted Docker deployment Bring Your Own AI Key No log-content storage by LogsFast Testing Audit Mode HTTP POST ingestion Versioned Docker updates

SIMPLE SELF-HOSTED SETUP

Get LogsFast Running in 3 Simple Steps

Deploy the Docker container inside your infrastructure, connect your monitoring source, and configure where qualifying events should go.

STEP 01

Configure Your Deployment

Create your config.json and provide your alert threshold and custom classification instructions. Set your license key, Typesafe API key, and destination webhook securely using environment variables or secret files.

STEP 02

Launch Your Docker Container

Run the LogsFast Docker image with your required environment variables, resource limits, and configuration. LogsFast starts its HTTP ingestion endpoint on port 8080 by default.

STEP 03

Test and Connect Your Alerts

Verify the health endpoint, send a sample alert, inspect the processing result, and connect your monitoring tool to the LogsFast HTTP endpoint.

Test in staging first. Validate masking, severity scores, threshold behavior, and webhook delivery before routing production traffic.

View Deployment Guide

PREDICTABLE LICENSING

Simple Pricing. Flexible Deployment.

Start with self-hosted AI log filtering using your own Typesafe API key. Enterprise capabilities are in development.

Early Access Offer: Free for the first 10 teams for 30 days in exchange for feedback. Just sign up — no credit card required.

Available now

Standard

Self-hosted log filtering for teams using their own AI scoring API key.

$199 / month
  • Docker container deployment
  • Bring Your Own Typesafe API key
  • Local sensitive-data masking
  • AI severity scoring (1–5)
  • Configurable alert thresholds
  • Custom classification instructions
  • Destination webhook routing
  • Health monitoring and usage tracking
  • No activation limit under Standard
Planned · Coming soon

Enterprise

Designed for organizations seeking a self-hosted AI scoring model without per-token AI provider costs.

Coming Soon
Planned capabilities
  • Self-hosted AI scoring model
  • No per-token external AI API cost
  • Log scoring within your infrastructure
  • Reduced dependency on external AI processing
  • Enterprise deployment options

Pricing note: Standard includes the LogsFast software license. Typesafe AI usage is billed separately by the provider under your own account. The Enterprise plan is planned and not currently available.

Got Questions?

Frequently Asked Questions

Everything you need to know about LogsFast, AI scoring, data handling, and deployment.