Debugging

Debug requests with full visibility

See every detail of your AI requests. Timeline views, full payloads, token usage, and error traces all in one place.

Request/Response logs Timeline views Failover tracking Event callbacks

Visual

Request logging flow

Every request is captured with full context from start to finish.

01

Client request

Payload + metadata captured

02

ModelRiver processes

Routing + failover logic

03

AI provider processes

Model generates response

04

Response returned

Data + token counts

05

Request log stored

Indexed, searchable, replayable

Timeline Views

Visualize request flow

See the full journey of your requests with interactive timelines. Track failover attempts, callback responses, and provider retries with visual indicators showing success, failure, and timing.

Primary request visualization
Failover attempt tracking
Event callback visualization

Full Payload Inspection

Deep data exploration

Inspect complete request and response payloads with syntax-highlighted JSON. Collapsible tree views make it easy to navigate complex nested structures and find exactly what you need.

Syntax-highlighted JSON
Collapsible tree viewer
Copy-paste ready format

Provider & Model Tracking

Know what ran where

Every log shows which provider and model handled the request. See backup attempts when primary providers fail. Track which models work best for your use cases.

Provider icons & labels
Model version tracking
Backup provider visibility

Token & Cost Analytics

Track usage & expenses

See exact token counts for every request: prompt tokens, completion tokens, and total usage. Calculate costs per request and track spending patterns across different models.

Prompt & completion tokens
Per-request cost breakdown
Historical usage trends

Advanced debugging capabilities

Production-grade tools for complex troubleshooting

Latency Tracking

Measure request duration down to the millisecond. Identify slow requests and optimize performance. Compare latency across providers and models.

Error Diagnostics

Get detailed error messages with full context. See which step failed in multi-provider workflows. Access provider-specific error codes and descriptions.

Customer Data Echoing

Cache custom fields from requests and see them in logs. Track user IDs, session IDs, and metadata. Connect logs to your application state for easier debugging.

Filtering & Search

Filter logs by status, provider, model, or workflow. Separate playground tests from production traffic. Search by channel ID, event name, or customer data.

Event Workflow Tracking

Follow event-driven workflows from AI generation to backend callback. See webhook payloads sent to your server. Inspect callback responses and timeout errors.

Real-time Refresh

Logs update automatically as requests come in. No need to manually refresh. See the latest data instantly while debugging active issues.

Debug faster, ship confidently

Real use cases from production teams

Production Issue Resolution

User reports an error? Search logs by their user ID in customer data. See the exact request payload, response, and error message. Reproduce the issue with the same inputs.

Failover Verification

Check if your backup providers are working correctly. Timeline view shows which providers were tried and why they failed. Ensure your redundancy strategy is solid.

Event Callback Debugging

Event-driven workflow not completing? See the webhook payload sent to your backend. Check if the callback was received. Identify timeout issues and fix your integration.

Cost Optimization

Analyze token usage per workflow. Find requests using excessive tokens. Switch to more efficient models based on actual usage data. Track cost savings over time.

Start debugging smarter

Every request is logged automatically. No setup required. Start seeing what's happening under the hood from your first API call.