Bronto for Claude

Ask Claude About Your Live Telemetry

Enable the Bronto connector and Claude can search your logs, traces, and metrics directly. Investigate incidents in plain language, backed by 12 months of always-hot, full-fidelity data.

1. Create a Bronto account

Try Bronto with a 2 week free trial

By signing up, you agree to our Terms of Service and Privacy Policy.

Already have an account? Sign in

2. Enable the Bronto connector in Claude

Find Bronto in the Claude connector directory and connect your account.

Claude · Bronto connector

Why did checkout error rates spike at 14:05?

bronto.search_logs · 1.2M events · 0.4s

Errors rose 12x at 14:05, all from payments-api. The top pattern is a connection timeout to the card processor, first seen in deploy a41f9c. Traces show p95 latency jumping from 180ms to 4.2s.

Summarize the top new log patterns since the last deploy.Which services had p95 latency over 500ms this week?

Built-in AI troubleshooting that slashes your MTTR

Claude queries Bronto through the MCP server, so answers come from your real data: logs, traces, and metrics correlated in one place, without jumping between tools.

Why Bronto + Claude

Give Claude the full context of your production systems

Bronto is a full-stack intelligent observability platform designed for both humans and agents. Claude, Claude Code, and any MCP-compatible tool can query and analyze your observability data instantly.

MCP server out of the box
Claude and Claude Code can search logs, traces, and metrics directly through Bronto's MCP server.
Cost-aware by design
Up to 100x cheaper than legacy hot tiers, so agents can query freely.
Context enrichment
Statement Patterns and Statement IDs group similar logs and link back to the exact code path, so Claude reasons over structure instead of raw text.
Structured, not stringy
Every field parsed at ingest, so tool calls return typed JSON instead of regex guesses.
OTEL GenAI + OpenLLMetry support
Monitor your own Claude-powered apps alongside the rest of your stack.

Logs

Metrics

Traces

Intelligence layer

Prompts to get you started

Ask Claude questions like these. Bronto's tools handle dataset discovery, schema lookup, and sub-second search behind the scenes.

Show me all error-level logs from the payments service in the last 15 minutes and group them by error message.

Are there any new error patterns in our application logs over the last 24 hours that weren't present yesterday?

Which datasets contain production CDN logs for our edge or delivery layer?

Which datasets ingested the most data in the last 7 days, and how does that compare to the previous week?

Find any IPs making more than 100 requests per minute in the last hour and show what they're hitting.

Show me the p50, p95, and p99 latency for the checkout endpoint over the past day.

Find any requests with response times over 2 seconds in the past hour and break them down by endpoint.

Show me the top 10 services by log volume over the last 24 hours.

Show authentication failures grouped by source IP and account over the last 6 hours.

Turn investigations into automations

Once Claude finds the query that exposes an issue, Bronto's MCP tools let it turn that work into monitors, saved searches, and coverage checks your team can reuse.

Turn investigations into monitors

Make sure a problem you just solved gets caught automatically next time.

create_monitor
Turns investigation queries into muted monitors, in bulk, with a dry run and an attached Investigation Prompt.
validate_monitors
Flags monitors that can no longer fire because data stopped flowing or they were changed.
search_monitors
Finds existing monitors and their trigger and recovery history.

Preserve analysis context

Keep the views an engineer built during an incident so the team can reuse them.

create_saved_search
Saves and tags queries as timeseries, top-lists, tables, or geomaps in the Bronto UI.
get_saved_searches
Shows what searches already exist, so agents reuse them instead of rebuilding one-offs.
run_saved_search
Runs a validated saved search with a new time window, saving tokens.

Coverage, usage, and health

See how much of your data is actively being searched or monitored.

get_coverage
Reports whether each dataset is monitored, searched, and still ingesting, to surface blind spots.

Read the full post: Turning Investigations into Automations with the Bronto MCP Server →

Watch the demo

Monitor Claude usage with traces

See how Bronto uses traces to show what Claude is doing, from each request to the tokens it uses.

FAQ

Common questions

Give Claude the context it needs

Sub-second search, 12 months hot retention, and an MCP server built for agents.

Book a Demo Start free