Recall — Structured Logging Built for AI

Before

How Claude searches logs today

C

Claude

Y

You

Why did the payment fail for user 12345 yesterday?

C

Claude

I don't have access to your logs.

Could you search your logging system for entries related to user 12345 and share what you find?

I'd need to see the request logs, payment gateway responses, and any error messages.

Y

You

*Opens Datadog, writes complex query, scrolls through thousands of lines...*

After

How Claude searches logs with Recall

C

Claude

Recall connected

Y

You

Why did the payment fail for user 12345 yesterday?

C

Claude

Let me search the logs for that user...

Using recall.query_logs

recall.query_logs("payment failed user:12345 yesterday")

Found 3 relevant entries

14:23:01 PaymentController#create started for user 12345

14:23:02 Stripe API call: card_declined, insufficient_funds

14:23:02 Payment failed, notifying user via email

Found it. The payment failed because Stripe returned card_declined with reason insufficient_funds.

The user was notified via email at 14:23. This wasn't a bug—their card simply didn't have enough funds.

Query time: 1.2 seconds

The Difference

Without Recall

— Claude can't search your logs
— You write complex query syntax
— Scroll through thousands of lines
— Context scattered across tools

With Recall

+ Claude searches logs directly
+ Ask questions in plain English
+ Get summarized, relevant results
+ Connected to errors, metrics, traces

How It Works

1

Your app logs normally

Use Rails.logger as usual. Recall captures and structures everything.

Rails.logger.info "Payment processed", user_id: 123, amount: 99.00

2

Logs become searchable

Structured JSON, indexed, and ready for natural language queries.

{ "message": "Payment processed", "user_id": 123, "amount": 99.00, "request_id": "abc-123" }

3

AI asks questions

Claude queries in plain English. Recall translates to the right search.

"Show me failed payments over $100 in the last hour"

4

Get instant answers

Relevant logs, summarized and contextualized.

Found 7 matching entries in 0.8s

The Tools AI Uses

Available via MCP (Model Context Protocol)

recall.query_logs()

"Search logs in plain English"

recall.get_request_trace(id)

"Show all logs for this request"

recall.get_patterns()

"Find recurring log patterns"

recall.summarize()

"Summarize logs for this period"

Get Started

Works with Rails.logger out of the box.

# Add to your Gemfile

gem 'brainzlab'

# That's it. Your logs are now searchable by AI.

# No changes to your logging code needed.

Try It Now

Run Recall locally in 5 minutes. Docker required.

Quickstart Guide Follow the Journey