ffmpeg API — media processing over HTTP

Run ffmpeg in the cloud without any infrastructure. Send a media URL to an endpoint and get the processed file back. Convert, compress, trim, make GIFs, grab thumbnails or extract audio — all pay-per-operation, no server to manage, no ffmpeg to install.

• ⚡ HTTP API (Standby): POST /<operation> with a JSON body, get a JSON reply with a link to the result
• 🔒 Safe by design: you send URLs and parameters, never raw shell commands
• 💵 Pay per operation, from $0.02
• 🤖 Callable from any language, from n8n/Make/Zapier, or from an AI agent

Endpoints

Base URL (Standby): https://synthetic-ia--ffmpeg-api.apify.actor with header Authorization: Bearer <APIFY_TOKEN>.

Endpoint	Body	Does
POST /convert	{ url, format: mp4|webm|mkv|mov }	Transcode to another container/codec
POST /compress	{ url, crf?: 18-40, preset?: ultrafast..slow }	Re-encode smaller (H.264)
POST /trim	{ url, start?, duration? }	Cut a clip (seconds or HH:MM:SS)
POST /gif	{ url, start?, duration?: 0.5-15, fps?: 4-30, width?: 120-1080 }	High-quality GIF from a clip
POST /thumbnail	{ url, time?, width?: 120-1920 }	Still JPG at a moment
POST /extract-audio	{ url, format: mp3|aac|m4a|wav|ogg|flac }	Pull the audio track
POST /advanced	{ inputs:[url,…], filterComplex?, map?, outputFormat, videoCodec?, audioCodec?, crf?, preset?, videoBitrate?, audioBitrate?, fps?, size?, pixFmt? }	Full ffmpeg: any filtergraph + codecs, safely validated
POST /async	{ operation, …params } → { runId }	Start a long job; then GET /status/<runId>
GET /status/<runId>	—	Poll an async job; returns the result when done

GET / returns this list as JSON.

Full power — /advanced

Expose (almost) everything ffmpeg can do through a validated interface: pass up to 4 inputs and a filterComplex graph (overlays, concat, scaling, color, speed, watermarks, hstack/vstack…), pick codecs/format/bitrate/fps/size. For safety, file-access filters (movie, subtitles, …) and extra inputs are rejected — you can't read the server's files.

curl -X POST "https://synthetic-ia--ffmpeg-api.apify.actor/advanced" \
  -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
  -d '{"url":"https://example.com/clip.mp4","filterComplex":"[0:v]scale=720:-1,hue=s=0[v]","map":["[v]"],"outputFormat":"mp4","crf":24}'

Long jobs — /async

Heavy transcodes can exceed the 4-minute synchronous window. Start them async and poll:

RID=$(curl -s -X POST ".../async" -H "Authorization: Bearer $APIFY_TOKEN" -d '{"operation":"convert","url":"…","format":"webm"}' | jq -r .runId)
curl ".../status/$RID" -H "Authorization: Bearer $APIFY_TOKEN"   # → { "status": "SUCCEEDED", "result": { "url": "…" } }

Response

{ "ok": true, "operation": "gif", "url": "https://api.apify.com/v2/key-value-stores/.../records/output/gif-....gif", "contentType": "image/gif", "sizeBytes": 812345, "tookMs": 1840 }

Example

curl -X POST "https://synthetic-ia--ffmpeg-api.apify.actor/gif" \
  -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
  -d '{"url":"https://example.com/clip.mp4","start":"0:03","duration":4,"width":480}'

Prefer a one-off run? Start the Actor normally with input { "operation": "compress", "url": "https://…", "crf": 30 }.

Limits

• Input max 250 MB, http(s) URLs only.
• Synchronous response under 4 minutes — great for clips, thumbnails, GIFs, conversions. For very long transcodes, run it as a normal (async) Actor run.
• No GPU (Apify is CPU-only), so no hardware encoding; H.264/VP9 software encoding is used.

Pricing

Per successful operation, from $0.02 on higher plans (platform usage included). You only pay when an operation succeeds.

FAQ

Can I send raw ffmpeg commands? No — for security you use the endpoints above with parameters. This covers the common operations safely. Need another operation exposed? Open an issue.

Where does the output go? To the run's key-value store; the response gives you a direct URL to download it.