Step 1: Generate an API key
In the Shilo dashboard, navigate to Settings > API Keys > Generate API Key and create a read-write key. Copy the key immediately, it is only shown once. Store it as a secret in your environment (for example,SHILO_API_KEY) and never commit it to source control.
Step 2: Prepare a recording URL
Host your call recording at a publicly accessible URL that Shilo can fetch over HTTPS. Common patterns include a signed S3 URL, a Cloudflare R2 public link, or any CDN-hosted audio file. Make sure the URL resolves without additional authentication for the duration of processing.Step 3: Submit the call
Send aPOST to https://api.shilo.ai/api/v1/calls with your API key in the x-api-key header and a JSON body containing every required field.
cURL
202 Accepted with the queued call event.
202 Accepted
Step 4: Poll for analysis
Save thecall_event_id and poll GET /api/v1/calls/call_event_id:{call_event_id}/analysis using the same API key. Continue polling until progress reaches 100.
cURL
progress: 100 means processing is complete. Individual analysis objects can still be absent or null depending on the call’s content and length, so treat each sub-object as optional in your integration.
Step 5: Inspect the analysis
Once processing is complete, read the fields your workflow needs. The analysis response can include:rating, a numeric quality score for the call.summary, an AI-generated recap.coaching, structured coaching feedback for the rep.disposition, the call outcome classification.objections, buyer objections detected in the conversation.action_items, follow-ups surfaced from the discussion.speakers, AI-resolved user and client names.
null on a given call, so check for presence before accessing nested fields.
For deeper coverage of every request and response field, see the Submit Calls and Retrieve Analysis guides.
Next Steps
Submit Calls
Learn all required and optional fields for the call creation endpoint.
Retrieve Analysis
Deep-dive into every analysis field and the polling pattern.
Identifiers
Understand the prefix:value identifier system used across all resources.
Pagination
Page through large result sets with cursor-based pagination.