REST API
Read your church’s Glossa data from your own software: which services exist, whether a service is live right now and who is listening, past sessions and their transcripts, billed usage, and the weekly recap materials. The API is available on every plan, including Pay As You Go and the free trial.
- Base URL:
https://api.glossa.live/v1/ - Reference: the API reference page, or the machine-readable OpenAPI document (OpenAPI 3.1) to open in Swagger UI or Postman, or to hand to an AI agent.
What you can do today
Section titled “What you can do today”- Check whether each of your services is live and how many people listen in each language, for a conference portal, a lobby screen or a monitoring script.
- Pull past sessions with their duration, listeners and billable minutes, and download transcripts as JSON, plain text or SubRip subtitles.
- Read billed minutes by day, language or service, and your plan’s remaining hours.
- Fetch your weekly recaps: title, big idea, quote, caption, scripture references, stats and the generated image and PDF files.
- Read the RTMP stream key of a service (opt-in scope) for hands-free sessions.
The first version is read-only. There is no endpoint to start or stop a session: for hands-free start and stop, send an RTMP feed to the service’s stream key, as described in RTMP ingest and stream keys. The session starts when your encoder connects and ends when it disconnects. To be told when a session starts or ends, who is listening, or what was said, register a webhook instead of polling.
Creating a key
Section titled “Creating a key”In the dashboard open Settings → API keys (glossa.live/account/settings/api-keys) and click Create key. Give it a name, choose what it may read and, optionally, when it expires (no expiration by default, or 7, 30 or 90 days, or 1 year). Then copy the key: it is shown once and never again.
Keys look like glk_live_ followed by 43 characters. They are secret: keep them on a server. Never put a key in a public web page or a mobile app. A church can have up to 10 active keys; create one key per integration so you can revoke them separately. Revoking a key takes effect within a minute, and an expired key gets 401 like a revoked one.
If you do not see the API keys page in your dashboard, write to info@glossa.live and we will issue a key for you.
Authentication
Section titled “Authentication”Send the key as a bearer token on every request:
curl -H "Authorization: Bearer glk_live_…" https://api.glossa.live/v1/meA missing or unknown key returns 401. A key without the needed scope returns 403.
If you call the API with Python’s built-in urllib, set your own User-Agent header. Its default, Python-urllib/3.x, gets a 403 from Cloudflare (error 1010) instead of a JSON API error. Libraries such as requests are not affected.
| Scope | What it unlocks |
|---|---|
services:read |
Services, live status and listeners |
sessions:read |
Sessions and transcripts |
usage:read |
Billed minutes and plan hours |
recaps:read |
Weekly recaps and their files |
stream_keys:read |
RTMP stream keys (sensitive, opt in when creating the key) |
Is my service live?
Section titled “Is my service live?”GET /v1/services/{id}/status answers from the live system. The service id is the UUID in your listener link (glossa.live/your-church/{id}); GET /v1/services lists them.
curl -H "Authorization: Bearer glk_live_…" \ https://api.glossa.live/v1/services/3f2b7c1e-9d4a-4b8e-a1c2-0d5e6f7a8b9c/status{ "service_id": "3f2b7c1e-9d4a-4b8e-a1c2-0d5e6f7a8b9c", "live": true, "basis": "memory", "source": "rtmp", "session_id": "5e6f7a8b-9c0d-4e1f-8a2b-3c4d5e6f7a8b", "started_at": "2026-10-25T10:02:14.000Z", "duration_seconds": 1745, "listeners": { "total": 38, "by_language": [ { "code": "es-ES", "count": 31 }, { "code": "pt-BR", "count": 7 } ] }, "as_of": "2026-10-25T10:31:19.000Z"}liveis true or false. While live,sourcesays where the audio comes from (microphone,rtmp,srtorurl), andsession_id,started_atandduration_secondsdescribe the running session.listeners.totalandlisteners.by_languagecount the people listening right now, caption overlays included.basisismemorywhen the answer comes from the live state. It isdatabasewhen only the stored flag says the service is live and listeners are unknown.as_ofis when the snapshot was taken. Poll every 5 seconds or slower.
Endpoints
Section titled “Endpoints”| Endpoint | Scope | Returns |
|---|---|---|
GET /v1/me |
any key | Church, plan and hours, pay-as-you-go balance, the key’s scopes |
GET /v1/services |
services:read |
Your services with listener, embed and iframe URLs |
GET /v1/services/{id} |
services:read |
One service, its session cap and whether the translation webhook is on |
GET /v1/services/{id}/status |
services:read |
Live state and listeners per language |
GET /v1/services/{id}/stream-key |
stream_keys:read |
The RTMP URL and stream key |
GET /v1/services/{id}/sessions |
sessions:read |
Past sessions, newest first |
GET /v1/sessions/{id} |
sessions:read |
One session with per-language minutes and listeners |
GET /v1/sessions/{id}/transcript |
sessions:read |
The transcript as JSON, plain text or SubRip |
GET /v1/usage |
usage:read |
Billed minutes and cost by day, language or service |
GET /v1/recaps |
recaps:read |
Weekly recaps, newest first |
GET /v1/recaps/{id} |
recaps:read |
One recap |
GET /v1/recaps/{id}/files/{key} |
recaps:read |
A 10-minute signed download link for a recap file |
Every response is JSON with snake_case fields and UTC timestamps in ISO 8601. Lists are paginated with page (from 1) and limit (default 20, at most 50) and carry a total.
Errors and limits
Section titled “Errors and limits”Errors come back in one shape:
{ "error": { "code": "not_found", "message": "Service not found", "request_id": "9b1c2d3e-…" }}| Status | code |
Meaning |
|---|---|---|
| 400 | invalid_request |
A parameter is malformed; details lists the issues |
| 401 | unauthenticated |
Missing, unknown, revoked or expired key |
| 403 | insufficient_scope |
The key lacks the scope the endpoint needs |
| 404 | not_found |
No such resource for your church |
| 429 | rate_limited |
Too many requests; wait for Retry-After seconds |
| 500 | internal |
Something failed on our side; quote the request_id |
Quote the request_id (also sent as the Glossa-Request-Id header on every response) when you write to support.
Each key may make 300 requests per minute, and each IP address 600 requests per minute across all endpoints and keys (requests with a missing or wrong key count too). Every response carries RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset; a 429 also carries Retry-After.
- Usage is billed per minute of spoken content per language; silence and songs are not billed.
cost_usdis null on a subscription. - A transcript only has text in a language for the minutes someone was listening in it.
- Status counts caption overlays (embed pages) as listeners too.
- Keys belong to the church, not to a person. Anyone who can open your Settings can create and revoke them.
- The API ships together with the Settings → API keys page. If
https://api.glossa.live/v1answers 404, that release has not reached production yet.