Debug Logs
Salesforce debug logs, explained by AI
Pull the org’s captured debug logs, read them in a viewer that makes the noise scannable, then select up to five related logs and have the AI correlate them — pinpointing exceptions, the slowest SOQL and DML, and governor-limit consumption.
What you can do
- The recent-log list with operation, user, status, duration, size, and time
- A raw-body viewer that makes a wall of log output scannable
- Select up to five related logs and analyse them together as one investigation
- The AI reads each log through its own tools — no copying and pasting into a chat box
- Exceptions pinpointed, the slowest SOQL and DML surfaced, governor-limit usage explained
- Ask follow-up questions in the same conversation, with your full org context available
- Log bodies are stored once pulled, so re-reading costs no API calls
How it works
Turn on a trace flag
Logs only exist while a trace flag is active. Set one in Setup → Debug Logs for the user you are chasing, then reproduce the behaviour.
Pull and read
The log list appears with operation, status, and duration. Open any one to read it properly.
Analyse together
Tick up to five related logs, add a question, and let the AI correlate the evidence across all of them.
Why it's different
It reads the logs, you ask the questions
Pasting a debug log into a chat window fails on the first big one — logs are enormous and the interesting part is rarely at the top. Here the AI reads each selected log through its own tools, correlates several from the same investigation, and has your schema, automation, and Apex source available while it reasons. You get the explanation, not a truncation.
Try asking
- “Why did this transaction hit the SOQL query limit?”
- “What threw the exception in this log, and what called it?”
- “These three logs are the same user retrying — what is failing?”
- “Which part of this transaction is actually slow?”
- Hand a log to the AI instead of reading 40,000 lines yourself
Debug Logs — questions
Works even better with
AI Org Chat
Ask your org anything. The assistant inspects schema, automation, code, and live data with two dozen investigation tools — and remembers what it learns.
Learn moreFlow Health
Which flows are failing, how often, and for whom — real failure rates from interview logs, trended over time.
Learn moreDevelopment Studio
An AI pair programmer that writes Apex, LWC, and Flows — then reads the real compile errors and test failures and fixes them.
Learn morePut Debug Logs to work on your org
Start free with the Metadata Explorer — every AI feature is free while the beta runs, and a competitive price is still being worked out.