# How Hedra uses Together AI

**60%**: “Hedra achieved 60% total cost reduction through Together AI's competitive pricing and performance improvements.”

As published on [together.ai](https://www.together.ai/customers/hedra). Captured by usedby on 2026-11-21.

- Company: [Hedra](https://www.usedby.ai/companies/hedra.md)
- Tool: [Together AI](https://www.usedby.ai/tools/together-ai.md)
- Industry: [Artificial Intelligence](https://www.usedby.ai/companies/industry/artificial-intelligence.md)

## The use case: how Hedra deploys Together AI

Hedra runs its video generation models on Together AI's GPU clusters for training and auto-scaling inference, using Dedicated Container Inference. Together AI's engineers also collaborate on kernel optimization to speed up the models, including porting them to Blackwell.

Summary written by usedby from the source page, in English. The figures are those of Together AI and Hedra, not ours.

> Hedra achieved 60% total cost reduction through Together AI's competitive pricing and performance improvements. This cost reduction, combined with halving GPU time per video over 12 weeks of optimization, directly enables competitive pricing while maintaining healthy unit economics for millions of video generations.

- **3x**: “Recent collaboration ported Hedra's models to Blackwell architecture, achieving 3x faster inference compared to previous hardware generations.”
- **300x**: “Since seed stage, the company's compute needs have grown 300x while maintaining consistent performance and cost efficiency at massive scale.”
- **20%**: “Michael can text Together AI's team for emergency GPU additions, with 20% cluster expansions completed within hours.”

> Training our omnimodal Character-3 model required infrastructure designed for large-scale AI. The Together Frontier AI Factory delivered the performance we needed to push the boundaries of multimodal video generation. Together AI understands what builders need — and that made all the difference.
>
> Michael Lingelbach, CEO and Founder

Source: [together.ai](https://www.together.ai/customers/hedra), captured 2026-10-03.

> Together AI’s infrastructure has the capacity to soak up our viral moments without breaking a sweat. During major traffic surges, Dedicated Container Inference scales seamlessly while maintaining performance. And because we trained on Together’s Accelerated Compute, deploying to production was frictionless—one platform, zero artifact transfers, no deployment headaches.
>
> Terrance Wang, Founding ML Engineer

Source: [together.ai](https://www.together.ai/customers), captured 2026-10-03.

## Other AI tools used by Hedra

- [ElevenLabs](https://www.usedby.ai/tools/elevenlabs.md) — Confirmed
- [Gemini](https://www.usedby.ai/tools/gemini.md) — Confirmed
- [Cartesia](https://www.usedby.ai/tools/cartesia.md) — Declared by the company
- [fal](https://www.usedby.ai/tools/fal.md) — Declared by the company
- [Featurebase](https://www.usedby.ai/tools/featurebase.md) — Observed
- [SigNoz](https://www.usedby.ai/tools/signoz.md) — Observed

## Together AI users in Artificial Intelligence

- [Arcee AI](https://www.usedby.ai/companies/arcee-ai.md) — Confirmed
- [Cartesia](https://www.usedby.ai/companies/cartesia.md) — Confirmed
- [Decagon](https://www.usedby.ai/companies/decagon.md) — Confirmed
- [Deep Cogito](https://www.usedby.ai/companies/deep-cogito.md) — Confirmed
- [Dippy AI](https://www.usedby.ai/companies/dippy-ai.md) — Confirmed
- [LegionEdge](https://www.usedby.ai/companies/legionedge.md) — Confirmed

## What usedby checked

We compared the story with its live page on 2026-11-21.

- Checked: the figure 60% is printed word for word on the page, near the name of Hedra.
- Checked: the passage quoted above is copied word for word from the page, near the name of Hedra.
- Checked: Hedra uses Together AI. Confirmed line. Latest check across sources: 2026-11-22.
- Not checked: the result itself. We quote it; we did not measure it.

---
Source: https://www.usedby.ai/case-studies/hedra-together-ai · How we check: https://www.usedby.ai/methodology
