Company profile · 3 min read
Cloudglue: video context for AI
Cloudglue turns video libraries into searchable, structured and citable context for AI products and agents.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
What it does
Cloudglue is a video context layer for AI. Its API and Tinycloud agent turn video libraries into searchable, structured context that models can query, cite and reason over. The YC profile describes video understanding APIs; the current homepage shows chatbot/RAG across videos, structured extraction, search and analysis of hundreds of videos.
The fit is a developer team with a large video corpus—screen recordings, sales calls, training footage or operational video—that wants more than transcription. Cloudglue is useful when the model needs to locate moments, extract consistent fields or answer with playable citations.
Why I’d look closer
The product surface is specific about the work: natural-language search across thousands of hours, questions grounded in video context, schemas for consistent extraction and aggregate reports. The homepage says two hours of video can be processed in three minutes and that responses work across hundreds of videos. Those are company-reported performance claims, not an independent throughput benchmark.
The founder context is relevant. The YC biographies describe Amy Xiao with ML and infrastructure work at Snapchat, AWS and Arize AI, plus math/physics training. The public profile also links Tinycloud as an open beta agent, giving builders a path to evaluate the product before an API integration.
What I’d ask
How are timestamps and citations preserved, what video formats and languages are supported, and how does extraction quality change with accents, overlap and domain vocabulary? I’d run a representative corpus, inspect false retrievals and schema consistency, and compare hosted analysis with a local pipeline. Pricing and full API limits were not exposed in the sources checked.
My editorial take
Cloudglue is a good fit for teams that need video to become queryable evidence rather than a pile of captions. The API-first framing is clear. The proof is retrieval precision and usable citations on the customer’s own footage.
Quick facts
| Field | Sourced detail |
|---|---|
| Buyer fit | Developers building video search, RAG or structured extraction |
| Product | Video context API, Tinycloud agent, search, inquiry and schema extraction |
| Performance claim | Two hours processed in three minutes; company-stated |
| Public pricing | Not exposed in the sources checked |
Sources checked
Checked 2026-09-19.
| Source | Used for |
|---|---|
| YC company profile | Product, founder and Tinycloud context |
| Cloudglue homepage | Current API/video workflow and performance claim |
Cohort context
Cloudglue is listed in Summer 2024. In our 2026-09-18 directory snapshot, 161 of 248 listed companies in that cohort have YC’s primary industry label B2B (64.9%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:16:55.915Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Observed |
| Pricing link | Observed |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗