Ittybit — Video, audio, and image processing for apps and agents

Get up and running in seconds

Add video, audio, and image processing to your app with a few lines of code.

npm i -g @ittybit/cli

brew install ittybit/tap/ittybit

curl -fsSL https://ittybit.com/install.sh | sh

Install ittybit from github.com/ittybit for your runtime.
Run `ittybit login` to create an account and get an API key.
Run `ittybit video -i https://ittyb.it/sample.mp4 --width 240` to process your first file.

ittybit/cli 1k license Apache 2.0

GPU Cloud

Process files 50x faster on our distributed GPU infrastructure.

Consistent API

Strong types and structured outputs across video, audio, and images.

Open Source

Fully open source. Process files locally forever free.

Built for Agents

Media processing designed for AI agents and autonomous workflows.

Your agent's video, image, and audio toolkit

Convert formats, resize media, reduce filesize, extract thumbnails, or create adaptive streams. All with an API that stays consistent across video, audio, and images in just about any format; even those janky framerate uploads from the user with the random 2013 Android. The same props work identically across CLI, SDKs, and REST; from local dev to production.

Tools

Description

Convert between formats to optimize for size, quality, or compatibility. Use modern codecs like av1 and opus for dramatic size reductions, or normalize to h264 and mp3 for universal playback.

Example

ittybit video \
  -i s3://bucket/input.mov \
  --format webm

const task = {
  kind: "video",
  input: "s3://bucket/input.mov",
  options: { format: "webm" },
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "video",
  "input": "s3://bucket/input.mov",
  "options": { "format": "webm" }
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit video \
  -i s3://bucket/input.mov \
  --quality high

const task = {
  kind: "video",
  input: "s3://bucket/input.mov",
  options: { quality: "high" },
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "video",
  "input": "s3://bucket/input.mov",
  "options": { "quality": "high" }
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit video \
  -i s3://bucket/input.mov \
  --width 1080 \
  --height 1920 \
  --fit cover

const task = {
  kind: "video",
  input: "s3://bucket/input.mov",
  options: { width: 1080, height: 1920, fit: "cover" },
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "video",
  "input": "s3://bucket/input.mov",
  "options": { "width": 1080, "height": 1920, "fit": "cover" }
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit video \
  -i s3://bucket/input.mov \
  --start 10 \
  --end 30

const task = {
  kind: "video",
  input: "s3://bucket/input.mov",
  options: { start: 10, end: 30 },
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "video",
  "input": "s3://bucket/input.mov",
  "options": { "start": 10, "end": 30 }
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit adaptive \
  -i s3://bucket/input.mov

const task = {
  kind: "adaptive_video",
  input: "s3://bucket/input.mov",
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "adaptive_video",
  "input": "s3://bucket/input.mov"
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit audio \
  -i s3://bucket/input.mov \
  --format mp3

const task = {
  kind: "audio",
  input: "s3://bucket/input.mov",
  options: { format: "mp3" },
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "audio",
  "input": "s3://bucket/input.mov",
  "options": { "format": "mp3" }
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit image \
  -i s3://bucket/input.mov \
  --start 5.0 \
  --format webp

const task = {
  kind: "image",
  input: "s3://bucket/input.mov",
  options: { start: 5.0, format: "webp" },
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "image",
  "input": "s3://bucket/input.mov",
  "options": { "start": 5.0, "format": "webp" }
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

ittybit probe \
  -i s3://bucket/input.mov

const task = {
  kind: "probe",
  input: "s3://bucket/input.mov",
};

await fetch("https://api.ittybit.com/jobs", {
  method: "POST",
  headers: {
    Authorization: "Bearer ittybit_...",
    "Content-Type": "application/json",
  },
  body: JSON.stringify(task),
});

TASK='{
  "kind": "probe",
  "input": "s3://bucket/input.mov"
}'

curl -X POST https://api.ittybit.com/jobs \
  -H "Authorization: Bearer ittybit_..." \
  -H "Content-Type: application/json" \
  -d "$TASK"

Input

{
  "object": "external_file",
  "url": "s3://bucket/input.mov"
}

Output

{
  "object": "file",
  "kind": "video",
  "format": "webm",
  "width": 1920,
  "height": 1080,
  "duration": 45200,
  "codecs": { "video": { "codec": "vp9" }, "audio": { "codec": "opus" }, "string": "vp9, opus" }
}

{
  "object": "file",
  "kind": "video",
  "format": "mp4",
  "width": 1920,
  "height": 1080,
  "duration": 45200,
  "filesize": 8400000,
  "codecs": { "video": { "codec": "h264" }, "audio": { "codec": "aac" }, "string": "h264, aac" }
}

{
  "object": "file",
  "kind": "video",
  "format": "mp4",
  "width": 1080,
  "height": 1920,
  "duration": 45200,
  "codecs": { "video": { "codec": "h264" }, "audio": { "codec": "aac" }, "string": "h264, aac" }
}

{
  "object": "file",
  "kind": "video",
  "format": "mp4",
  "width": 1920,
  "height": 1080,
  "duration": 20000,
  "codecs": { "video": { "codec": "h264" }, "audio": { "codec": "aac" }, "string": "h264, aac" }
}

{
  "object": "file",
  "kind": "adaptive_video",
  "format": "m3u8",
  "variants": [
    { "width": 1920, "bitrate": 5000000 },
    { "width": 1280, "bitrate": 2500000 },
    { "width": 640, "bitrate": 1000000 }
  ]
}

{
  "object": "file",
  "kind": "audio",
  "format": "mp3",
  "duration": 45200,
  "filesize": 3600000,
  "codecs": { "audio": { "codec": "mp3" }, "string": "mp3" }
}

{
  "object": "file",
  "kind": "image",
  "format": "webp",
  "width": 1920,
  "height": 1080,
  "filesize": 184000,
  "codecs": { "image": { "codec": "webp" } }
}

{
  "object": "file",
  "kind": "video",
  "format": "mov",
  "width": 1920,
  "height": 1080,
  "duration": 45200,
  "fps": 23.976,
  "bitrate": 8500000,
  "filesize": 48062500,
  "codecs": { "video": { "codec": "h264", "profile": "High", "level": 4.1, "bitdepth": 8 }, "audio": { "codec": "aac", "sample": 48000, "channels": 2 }, "string": "h264, aac" }
}

Process any media file

Videos

h264 h265 vp9 av1 mp4 webm

Images

jpeg png webp gif heic avif

Audio

wav mp3 aac ogg flac opus

Local Files

Process files directly from your local filesystem. Great for development, CI/CD pipelines, and batch processing on your own infrastructure.

S3 Storage

Read inputs and write outputs directly from your own bucket. Use AWS, GCS, R2, Supabase, MinIO, or any other S3-compatible provider.

Public URLs

Pass any publicly accessible URL as input. Ittybit will fetch the file, check it's compatible, process it, and return the result. No need to download first.

Signed URLs

Use pre-signed or time-limited URLs for private files. Supports S3 signed URLs, GCS signed URLs, or any auth-gated endpoint accessible via GET request.

Offload heavy processing

When you're ready to scale past your demo, ittybit's cloud gives you reliable on-demand processing with a single line-of-code. Process parallel files at lightning-fast speeds on our custom GPU infrastructure. Skip terraform and provisioning. It scales up for millions of files when you need and down to zero when you don't.

22 min 4K → H264

Ittybit Cloud

48s

AWS MediaConvert ¹

5m 12s

FFmpeg ²

14m 25s

44 min 4K → AV1

Ittybit Cloud

7m 56s

AWS MediaConvert

3h 42m

FFmpeg

5h 16m

2hr lecture → 1080p ABR ladder ³

Ittybit Cloud

53s

AWS MediaConvert

32m

FFmpeg

1h 37m

Auto-scaling

0succeeded 0failed

Events

Parallel Processing

Submit multiple files at once and process them simultaneously across distributed workers. No shared state, no coordination overhead, no bottlenecks.

Scale to Millions

We handle the queues, provisioning, health checks, and capacity planning. You send files, we process them. Whether it's ten or ten million.

Scale to Zero

Pay only for the files you process. No idle instances, no reserved capacity, no fixed infrastructure costs. Usage scales linearly with your actual workload.

Automatic Retries

Transient failures are retried automatically with exponential backoff. Unrecoverable errors return clear debug hints, not cryptic exit codes.

Built for, and by, human developers

Our founding team has 35 years experience building video and imaging pipelines. All the parts that suck? We've battled through them multiple times. Now at ittybit we sweat the details for you. (So you can focus on all the other parts of your app that suck).

+30 -4,242

42 - ## TODO: Build Video Pipeline

43 - - [ x ] Shitty vibecoded proof-of-concept

44 - - [ ] Task queue: SQS, Cloud Tasks, or Redis?

45 - - [ ] Worker fleet: auto-scaling, health checks, draining?

46 - - [ ] GPU instances: spot pricing, availability zones?

47 - - [ ] FFmpeg builds: codec licensing, version pinning?

48 - - [ ] Storage layer: temp files, signed URLs, cleanup?

49 - - [ ] Retry logic: chunk failures, timeouts, DLQs?

50 - - [ ] Progress tracking: status polling, event streaming?

51 - - [ ] Monitoring: per-job logs, error rates, alerts?

42 + ## DONE: Build Video Pipeline

43 + - [x] Install ittybit CLI

44 + - [x] Run `ittybit video -i input.mov --quality high`

Resources

CLI API SDKs Docs

Works with all your existing tools

Models

Hosting

Workflows

Loved by AI agents

Get detailed metadata for every file, without parsing walls of stdout. Track granular progress without accidentally starting a giant infrastructure project. Or fire-and-forget with async tasks and reliable event-driven notifications.

Trigger

s3:ObjectCreated

bucket: podcasts

Agent

analyze_content()

Clip 1

video

start: 0, end: 12

Thumbnail 1

image

format: webp

Clip 2

video

start: 34, end: 48

Thumbnail 2

image

format: webp

Upload

R2 Connection

bucket: clips

Trigger

s3:ObjectCreated

bucket: podcasts

Agent

analyze_content()

Clip 1

video

start: 0, end: 12

Thumbnail 1

image

format: webp

Clip 2

video

start: 34, end: 48

Thumbnail 2

image

format: webp

Upload

R2 Connection

bucket: clips

Fully Typed

Zod schemas and TypeScript types for every request and response. Agents can validate parameters before sending, so errors surface at the schema level instead of mid-encode.

Structured Data

Responses, logs, events, and errors are all typed JSONL. Agents can parse every field directly without regex or string wrangling.

Pipe In/Out

Accepts JSON in, returns JSON out. Slot ittybit into any chain of CLI tools or agent steps. Compose it the same way you would jq, curl, or any unix tool.

Webhooks

Get async notifications when tasks complete. Ideal for event-driven architectures and set-and-forget workflows where polling is wasteful.

Add scalable media processing to your app in seconds

Create a free account, install the CLI, and process your first file in under a minute. No credit card, sales calls, or 90-page procurement doc required.

Create Account

View Docs Contact Us

p.s. There's no lock-in

The CLI and encoding engine are open source under Apache 2.0. Run locally in dev, self-host in production, or use our managed cloud. Switch between them with a single flag. We want you to stay because we've built the best system and we've earned your business every month, not because of some long-term contract bullshit or enterprise deal rug-pull. If you do leave, your code and your files stay yours.

ittybit/cli 1k license Apache 2.0

Run Locally

Process files on your own machine with zero network overhead. Great for development, testing, and CI/CD pipelines.

Self-Hostable

Deploy on your own infrastructure. Any server, any container service, any cloud provider.

Forever Free

The CLI will always be free and open source. No trial periods, no usage caps, no surprise invoices.

No Restrictions

No filesize limits, no format restrictions, no duration caps, no watermarks, no feature gates.

Stop FFighting with FFmpeg

Get up and running in seconds

Your agent's video, image, and audio toolkit

Process any media file

Videos

Images

Audio

Offload heavy processing

Built for, and by, human developers

Works with all your existing tools

Loved by AI agents

Add scalable media processing to your app in seconds

p.s. There's no lock-in