Getting started
What does Masana do?
We collect custom real-world data for AI and robotics teams: first-person and task video, field and street video, speech, images, and text. We also help owners of video, audio, and text archives license them to AI companies.
Who do you work with?
AI and robotics teams that need training or evaluation data, data companies that want a collection partner under their own brand, and organizations with archives to license.
How do we get started?
Send us your spec through the contact form. A rough brief is fine. We reply with a scope, a timeline, and a pilot proposal within two business days.
Do you have a spec template?
Yes. Our guide How to write a data collection spec includes a template you can copy.
Data & collection
What kinds of data can you collect?
First-person and task video, field and street video, speech and audio, images, and text and handwriting. If it can be filmed, recorded, photographed, or written by real people, we can scope it. See Solutions.
Where do you collect?
Anywhere you need it. Our home base is Bali, and our deepest network is across Indonesia and Southeast Asia. Projects elsewhere are scoped one by one, through our own contributors or local partners. If we can't do it well, we'll tell you up front.
Can you work to our existing capture guidelines?
Yes. Send them with your spec and we'll brief contributors to them exactly.
What formats do you deliver?
MP4 or WAV with JSON metadata and a CSV manifest by default, or RLDS and LeRobot formats on request. Folder structure and naming follow your spec.
How is data delivered?
Through a secure cloud transfer in the format and folder structure you specify, with metadata and labels included.
Pilots & pricing
How small can a pilot be?
Small enough to judge quality quickly. Typically a few hours of video or audio, or a few hundred images or text items. We'll suggest a size when we scope your project.
How long does a pilot take?
Usually 1 to 3 weeks from an agreed spec, depending on the task, location, and contributor profile.
How is it priced?
Every project is quoted up front: per approved hour of video or audio, per image, or per text item. Price depends on task, location, contributor requirements, and labeling.
Do we pay for files that fail review?
No. You only pay for data that meets the agreed spec. Files that fail are re-captured, not passed along.
Quality
How do you check quality?
Contributors get a written capture guide and a short trial first. Every file is then checked against your spec, and labels are reviewed by a second person before delivery. More on quality checks.
What happens if we're not happy with the pilot?
You give feedback, we update the capture guide, and the next batch reflects it. That loop is what the pilot is for.
Consent & legal
How do you handle consent?
Every contributor signs a release in their own language before anything is recorded, aligned with the data protection law where the data is collected, such as GDPR, CCPA, or Indonesia's UU PDP. See Consent & data ethics.
Can we see the consent records?
Yes. Consent records are available to you on request, and each batch ships with documentation of what was collected, where, when, and under which consent version.
Is the data exclusive to us?
Custom collections are made for your project and are not resold unless agreed. Exclusivity terms are agreed in writing before we start.
Do you record children or bystanders?
We never collect data from or about minors. Bystander faces in street and field footage are blurred before delivery, and recognizable people need their own signed release.
Archive licensing
How much is my archive worth?
It depends on volume, quality, language, uniqueness, and current buyer demand. We'll give you a realistic view after the rights review, before you sign anything.
Do I lose the rights to my content?
No. A license gives a buyer permission to use the material for the purposes agreed in the contract. You keep ownership.
What does it cost me?
Nothing up front. We earn a share of each deal, agreed in writing before any material is offered.
Our archive isn't organized. Is that a problem?
Not necessarily. Many archives are scattered across drives and tapes. We'll assess what's there and how much preparation it needs. More on archive licensing.
White-label partners
Can you collect under our brand?
Yes. We capture data under your brand and to your QA process, so it slots into what you deliver to your client.
Will you contact our clients?
No. You keep the client relationship. We work through you.
How do we start a partnership?
Tell us about your clients' needs on the contact form and choose "A data company" as your company type.
For contributors
How do I become a contributor?
Apply on the contributors page. We'll contact you when a project matches your location, languages, and equipment. You must be 18 or older.
How do I get paid?
You're paid for every file that passes review, at the rate agreed before you start.
What will my data be used for?
Training and evaluating AI systems. Every brief explains what you're recording and who it's for before you sign anything.
No questions match that search. Try another word, or ask us directly.
Still have a question?
Send it over. We reply within two business days.
Get in touch