Decrackle
Decrackle Inc. · Video · Other
Decrackle combines a browser playground with a REST API and TypeScript SDK for audio denoising and speech to text, so you can test one messy file and then run the same job in code.
About Decrackle
What Is Decrackle
Decrackle is an audio AI platform from Decrackle Inc., a Delaware company that sells three things: a content creator app, a conversation intelligence suite, and the API layer this page covers. The API handles two jobs. It runs AI noise removal on a recording, and it turns speech into text. The playground at studio.decrackle.io gives both jobs a point and click front end, which is handy when you want to know whether a file is salvageable before you write any code.
The pitch is simple. Field recordings, phone interviews, and podcast tapes often carry hiss, hum, and room echo that make them hard to listen to or to transcribe. Decrackle runs those files through its own models and returns a cleaner track or a text transcript.
The limits matter as much as the features. The published docs cover batch work only, with streaming transcription marked as coming soon. The same docs list eight supported languages for speech to text, while the marketing site mentions automatic detection across 30 or more. Treat the documented eight as the safe number for planning.
Getting Started
- Create an account at studio.decrackle.io, then open the playground from the left menu.
- Upload a sample file, pick denoise or transcribe, and let the job run. The result plays back in the browser, or shows up as text you can copy.
- Generate an API key in the dashboard. The free plan gives you a working key, so you can test the real endpoint before paying.
- Install the SDK with
npm install @decrackle/sdk, then pass your key through theDECRACKLE_API_KEYenvironment variable. - Call
client.denoise.process()orclient.asr.transcribe()and readoutputUrlortranscriptionTextfrom the response.
Product Information
A quick look at Decrackle's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Podcast and video editors
- Developers building media features
- Journalists and researchers
- Support and sales teams
Tasks
- Rescuing a recording with steady hiss or air conditioner hum
- Turning interviews into text
- Comparing model output before committing
- Adding cleanup to an existing pipeline
Scenarios
- One important interview sounds unusable and you need to know today whether it can be saved.
- A content team publishes weekly video and wants a repeatable noise removal step instead of manual repair.
- A product team is scoping an audio feature and needs a working prototype with pricing it can put in a budget.
Key features
Browser Playground for Testing Files
The Decrackle playground is the fastest way to see what the models do with your audio. You upload a file, choose denoising or transcription, and get output in the same window. No install, no key, no code. For teams evaluating an audio cleanup API, that removes the usual step of writing a test script just to hear a result.
Audio Denoising with EnhanceNet Models
Denoising is the core of the product. Decrackle runs files through its EnhanceNet family of models and returns a cleaned track. Free accounts get the v2 model only, while Professional unlocks the full model set and lets you pin versions so output stays consistent between releases. If you process client work, version pinning is the difference between a repeatable result and a surprise.
Batch Speech to Text with Language Detection
The transcription API takes a file and returns text, a detected language, and processing time. You can pass an ISO code such as en or ja when you already know the language, or leave the option out and let the service decide. The docs list English, German, French, Spanish, Italian, Hindi, Japanese, and Arabic. Paid accounts also return a quality score, which helps when you route low confidence transcripts to a human reviewer.
TypeScript SDK and REST API
The official SDK ships with full types, no runtime dependencies, and support for Node.js 18 and newer plus modern browsers. It handles automatic retries, configurable timeouts, request cancellation, and typed error classes, so a failed call surfaces as a specific error instead of a generic 500. Teams that prefer raw HTTP can call the REST endpoints directly and get the same jobs.
Job History and File Management
Every transcription and denoising run becomes a job you can list, fetch, or delete. Presigned download links expire, and the SDK includes a refresh call for that case, which matters for long lived pipelines that store URLs in a database. Storage quotas are visible per plan: 10MB on Free, 100MB on Professional, unlimited on Enterprise.
Plan-Based Limits and Output Formats
Free accounts get 30 minutes of processing a month, 10MB per file, 5 minutes per clip, and MP3 output. Professional raises that to 10,000 minutes a month, 100MB files, 120 minute clips, and adds WAV, FLAC, and Opus. The gap between 30 minutes and 10,000 minutes is the real dividing line, not the model list.
Pros and cons
Pros
- The free plan includes real API access, not just a demo page, so you can test your own files at no cost.
- The playground and the API run the same models, which shortens the path from "does this work" to "ship it."
- The TypeScript SDK is typed end to end and covers retries, timeouts, and error classes without extra libraries.
- Professional adds version pinning and four output formats, both useful when output feeds a production pipeline.
Cons
- Free accounts stop at 30 minutes a month and 5 minutes per file, so most podcasts will need a paid plan after the first test.
- The docs list eight transcription languages, well short of the 30 or more the marketing site suggests, so check your language before building.
- Streaming transcription is still marked coming soon, which rules out live captioning and real time call analysis for now.
Frequently asked questions
Decrackle cleans up noisy audio and turns speech into text. The playground is for quick tests on a single file, and the API is for running the same jobs inside an app, a publishing workflow, or a call analysis pipeline.
Related content
Explore related tools, skills, and articles for Decrackle.
Decrackle Alternatives
ngram 2.0
ngram · Video · Marketingngram 2.0 is an AI video generator for teams. It turns a prompt, a screen recording, or a product URL into a narrated explainer, product demo, or AI avatar video, and every scene stays editable.
OpenScreen 2
OpenScreen (open-source project) · Video · ProductivityOpenScreen 2 is a free, open-source screen recorder and demo video editor for Windows, macOS, and Linux, with automatic zooms, styled frames, local captions, and agent edits.
Skreno
Skreno · Video · ProductivitySkreno is a browser screen recorder that captures a tab, a window or your whole display, then attaches AI transcripts, summaries and console logs to one share link.