Tigris Data

@mixpeek t.ly/eUFPw

@mixpeek Files hit Tigris, Mixpeek's platform instantly indexes the frames, and you get a "searchable brain" with zero egress fees. See how:

Combing through 50 hours of videos for "penalty shot" is tough manually. With @TigrisData + @Mixpeek, your storage actually thinks.

I started building this as a side project while working full-time as CEO of Tigris. The goal: make training directly from object storage fast, operationally simpler, and without paying another vendor $$$ for data they already pay to store. Details here: tigrisdata.com/blog/training-…

If you're running large-scale training and want to simplify your pipeline, this is worth a look. We’re opening early access to TAG—happy to chat. 👉 Full benchmark: tigrisdata.com/blog/training-…

A simpler architecture works: Object storage + caching → direct training → full GPU utilization → less infra + less ops

💡 What this means for training You don’t need: - Local data staging - Complex preprocessing pipelines - Dataset sharding (for most workloads)

We benchmarked training directly from object storage—with no data hydrated to local storage Results were… not what most people expect 👇

@levelsio @NaourassD Have a look at @TigrisData as well. It’s S3-compatible.

Most AI/data pipelines are still batch-first: data lands → pipeline runs → downstream updates But what if data systems were reactive instead? On @pixeltablehq, you define transformations as computed columns so updates propagate automatically, without pipeline triggers and DAGs. It's simpler, and it's backed by Tigris object storage for all your multimodal data. Read more about reactive data models for AI workflows 👇 tigrisdata.com/blog/pixeltabl…

Today we're launching the Tigris Partner Integration. It's the fastest way for platforms to offer storage to their users. One API call. Isolated tenants. Globally distributed. No limits. Railway, Fly.io, Agentuity are already live on it. tigrisdata.com/blog/partner-i…

Need to provision buckets for your users? Our partner program is built for multi-tenancy from the start. And there are no regions nor account quotas to design around. Check out who's building Tigris buckets into their platforms. It feels so native, you'll never know unless you read the blog tigrisdata.com/blog/partner-i…

Startup life is stressful. One thing that's made it easier - the @TigrisData team. @ovaistariq and crew have been solid partners as we've built on their object storage. If you need S3-compatible storage for your next project, highly recommended!

It's been fun watching @parseablehq build an observability data lake on top of Tigris. They are basically enabling "AI-grade observability": model behavior, quality signals, token usage, and infra telemetry all in one place: tigrisdata.com/blog/case-stud…