Parasail
Per-token inference API and dedicated GPU endpoints for running open-weight AI models for AI-native startups
Highlights
- Hiring2 open roles
- New on EvolcAdded 2026-10-16
- Remote-friendly2 of 2 open roles remote or hybrid
About Parasail
Parasail runs an inference cloud that serves open-weight and frontier AI models through one OpenAI-compatible API. The home page describes a library of 40+ open and frontier models, and the pricing page cites access to 2M+ open models on its serverless endpoints.
The company targets AI-native startups and teams that run inference in production. Its pricing includes pay-as-you-go serverless endpoints billed per token, elastic endpoints priced at 1.25 times the model's serverless token rates, reserved dedicated deployments with negotiated SLAs, and batch jobs at the lowest per-token rate.
The website names Mike Henry, co-founder and CEO and former chief product officer of Groq, and Tim Harris, co-founder, as its founders. Open job listings include roles in San Mateo and San Francisco, plus Europe-remote roles in site reliability engineering. The site does not state a founding year or headcount.
Product: Serverless per-token API access to open-weight models, elastic and reserved dedicated GPU endpoints, and batch inference.
Open jobs (2)
Senior Site Reliability Engineer, Site Lead
New on EvolcRemote-friendlySenior Site Reliability Engineer
New on EvolcRemote-friendly
Leadership
- Tim Harris · Co-FounderSource: company website
- Mike Henry · Co-Founder & CEOSource: company website
Tech stack
- AWS
- Ashby
- Astro
- Google Analytics 4
- LinkedIn Insight
- Microsoft Clarity
- Plausible
- Sanity
- Vercel