Log Thread maps your repository locally, connects to your existing observability stack, and correlates runtime failures back to the exact files, functions, dependencies, and downstream impact. Your team gets a fast local Web UI, CLI workflows, and optional MCP-powered IDE assistance.
Works in the Log Thread Web UI and CLI, and also plugs into Cursor, Windsurf, Claude Code, and other MCP-compatible tools.
Issue: Checkout latency spiked after deploy
✦ Log Thread: Impact path touches checkout_service.py
✦ Related symbols: payment retry helper, cart serializer, order repository
✦ Downstream impact: billing API, queue worker, order sync job
Next: open the runtime panel to correlate logs and traces
$ logthread logs query prod-loki '{service="checkout"} |= "timeout"'
Events: 148 · Patterns: 5 · Sequences: 2
Top answer: timeout burst maps to reserve_inventory()
Evidence: retries began after dependency call to inventory gateway
Suggested follow-up: inspect retry loop and queue backpressure paths
Start with the local CLI, analyze your repository, then connect runtime providers such as Loki, Datadog, Splunk, Sumo Logic, Elasticsearch, CloudWatch, or New Relic.
Run pip install logthread
Use logthread login and logthread generate .
Add a provider with logthread logs connect ... or in the local Web UI
Use logthread ui, the dashboard, CLI analysis, or MCP-assisted editing
Log Thread is designed to sit on top of your current stack, not replace it. Point it at your existing observability systems and correlate production evidence back to the code graph.
Connect Grafana Loki, Datadog Logs, Splunk, or Sumo Logic with provider-specific credentials and saved defaults.
logthread logs connect --provider loki --name prod
Use Elasticsearch, CloudWatch, and New Relic through the same local runtime panel and query workflows.
Great for teams that already have governed logging and security controls in place.
Attach request IDs, trace IDs, service names, and other hints so Log Thread can cluster failures into actionable root-cause paths.
No need to duplicate logs or move your code into the cloud.
Individual users start with a personal workspace. Enterprise access is organization-based: your admin sets up a shared organization, verifies the company domain, enables SSO, and assigns seats.
Create an account, verify your email, subscribe, then run
logthread login,
logthread generate .,
and logthread ui.
Best for single users and small teams getting started quickly.
Enterprise access is not a per-user flag. It is a company organization with verified email domains, configured SAML or OIDC, and seat-managed memberships. Team members simply enter their work email at sign-in and Log Thread routes them to the right identity provider.
Use this for Okta and other SAML/OIDC-backed rollouts with centralized access control.
Trace production failures from log lines and request sequences back to the functions and modules that actually changed.
Graph context helps engineers and AI assistants find the right helpers, interfaces, and dependency boundaries instead of re-implementing them.
Impact analysis shows the APIs, workers, tables, and downstream workflows touched by a proposed fix before it ships.
Investigate from the local Web UI, run CLI analysis in the terminal, or surface the same context through MCP-enabled editors.
New teammates can get from install to first indexed repository and connected log source in a few guided steps.
Code analysis stays on the machine, while cloud features focus on auth, seats, billing, and team access control.
One plan for code intelligence, runtime analysis, team auth, and enterprise rollout.