Speak for developers
Turn audio and video into timed transcripts, styled captions, and platform-ready exports through the API or MCP.
API, SDKs, and MCP are in beta and open to every account.
# 1. Create a project from a media URL
curl -X POST https://capxion.me/api/v1/projects \
-H "Authorization: Bearer $SPEAK_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "source_url": "https://example.com/interview.mp4",
"kind": "video", "language": "auto" }'
# 2. Fetch the word-timed transcript once it's ready
curl https://capxion.me/api/v1/projects/$PROJECT_ID/transcript \
-H "Authorization: Bearer $SPEAK_API_KEY"
# 3. Export for two platforms with burned-in captions
curl -X POST https://capxion.me/api/v1/projects/$PROJECT_ID/exports \
-H "Authorization: Bearer $SPEAK_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "presets": ["youtube", "instagram_reels"],
"burn_captions": true }'{ "text": "Every",
"start": 0.42,
"end": 0.71 }00:00.420 --> 00:01.580 Every word counts
A simple workflow from upload to export.
Upload or reference your recording.
POST /api/v1/projects
{ "source_url": "https://…/talk.mp4",
"kind": "speech" }Review words and timing before export.
PUT /api/v1/projects/{id}/transcript
{ "lines": [{ "words": [{ "text": "hello",
"start": 0.42, "end": 0.71 }] }] }Track jobs and download completed files.
POST /api/v1/projects/{id}/exports
{ "presets": ["tiktok"],
"range_start": 12, "range_end": 42 }Same powerful transcription and captioning engine. Choose the integration that fits your workflow.
Bring caption workflows into your product.
{
"id": "3f0c7a52-8d7e-4a51-9b1e-2f1c0d6e8a41",
"title": "Interview",
"status": "ready",
"kind": "video",
"source": "video",
"language": "en",
"duration": 184.6,
"format": "9:16",
"style": { "template": "clean", "placement": "bottom" },
"error": null,
"job_id": "c5d1e9f0-1a2b-4c3d-8e9f-0a1b2c3d4e5f",
"created_at": "2026-10-10T14:01:07Z"
}Give your AI agent a captioning toolkit.
Caption this interview for Reels and YouTube.
Use typed clients for authentication, retries, and polling. Projects hold your media; transcripts contain word timings; exports create platform files; webhooks report job changes.
npm install @speak/sdkpip install speak-sdkSubmit work, track progress, and retrieve the result when it’s ready.
Completion event · example
{
"id": "6c0f2b9e-4d1a-4f3e-9a57-0e8b1c2d3f40",
"type": "export.completed",
"created_at": "2026-10-10T14:03:22Z",
"data": {
"project_id": "3f0c7a52-8d7e-4a51-9b1e-2f1c0d6e8a41",
"export_id": "b81e2d4c-77a0-4c0f-a6f3-5d9e1c2b3a70",
"status": "completed",
"error": null
}
}Transcripts, captions, and exports formatted for real-world use.
Get time-aligned transcripts with word-level timestamps.
{
"lines": [{
"words": [{
"text": "Every",
"start": 0.42,
"end": 0.71
}, …]
}]
}Generate beautifully styled captions, ready to share.
Every idea
starts with a conversation.
Export SRT and VTT subtitles for easy use anywhere.
1
00:00:00,420 --> 00:00:01,580
Every word counts
2
00:00:01,900 --> 00:00:03,240
so say what you meanRender for the formats your audience is on.
WEBVTT
00:00:00.420 --> 00:00:01.580
Every word counts
00:00:01.900 --> 00:00:03.240
so say what you meanGet from idea to finished captions in minutes.
Turn audio and video into transcripts, captions, and exports through the API or MCP.