LinkedIn · Aug 2024
Oscar Rovira announced a partnership with Black Forest Labs, saying Flux models would be commercially accessible on Mystic AI.
Low latency API to run and deploy ML models
Company research
Public company, workplace, funding, and market signals
Overview
Mystic (mystic.ai, also seen as Neuro AI) is a small UK-based MLOps / machine-learning infrastructure company that helps teams deploy and run ML models through serverless APIs and a Bring-Your-Own-Cloud (BYOC) platform, with a strong emphasis on low latency, GPU optimization, and reducing cold starts.
Primary product
Enterprise ML deployment and inference platform (serverless API + BYOC) for running custom or open-source ML models with GPU-optimized infrastructure.
Founded
2019
Headquarters
London, United Kingdom
Team size
1-10
Work style
Onsite
Industry
Software Development
Sub-industry
MLOps / machine learning infrastructure / AI inference platform
Offices
London, United Kingdom (HQ)
Bath, Bath and North East Somerset, United Kingdom
Business model
B2B SaaS
subscription
usage-based infrastructure
cloud pass-through / BYOC
Funding
Stage
Seed
Total raised
$7.3M
Latest round
Seed · Apr 2022
People & investors
Work & culture
Highly technical, entrepreneurial, early-stage startup culture with small-team ownership, collaborative product building, and an in-person office environment in Bath.
Work style
Onsite
Compensation
Public hiring signals include a senior software engineer salary band of £70,000-£110,000, plus equity compensation and competitive salary framing.
Benefits
Competitive salary
Equity compensation
Cycle-to-work scheme
Generous equipment budget
Flexible work-hours
High ownership / early-stage impact
Learning and development flexibility
Product & market
Pricing
Free-to-start serverless pay-per-millisecond API; BYOC runs on the customer’s cloud at provider cost; Turbo Registry plans were advertised at roughly $30-$100 depending on scale; team/project billing is described as standard SaaS billing rather than usage-based markup.
Differentiators
- —Deploy on your own cloud while keeping a serverless experience
- —Use cloud credits and pay provider cost in BYOC
- —Turbo Registry claims up to 15x faster model loading / up to 90% lower cold starts
- —Open-source Python SDK with direct access to the compute engine
- —Fractional GPU sharing and GPU optimization
- —Team-aware billing, roles, and shared BYOC clusters
- —One-click and no-code deployment flows
Technology
Python SDK
Kubernetes
containerd
Docker
Rust
NVIDIA MIG
OAuth / Google sign-in
GPU sharing
serverless GPU inference
Poetry
Pytest
pre-commit
Discord support/community
BYOC cloud integrations
Customers
Data scientists
Software developers
Enterprise ML teams
Teams deploying custom or open-source ML models
Users running workloads on AWS, Azure, or GCP
Competitors
Hugging Face
Replicate
Flower
TrueFoundry
Modelbit
MLEM
Giskard
Utilidata
Mind Foundry
InsightFinder AI
Scale & contact
Estimated revenue
≈$1.8M ARR in 2024 (third-party estimate); other public directories place it in the $1M-$10M revenue band.
Estimated monthly visits
18.9K
Traffic estimate as of Apr 2025
Latest news
LinkedIn · Aug 2024
Oscar Rovira highlighted the Turbo Registry, a Rust-based Docker registry that loads a 13GB image in 12 seconds and improves cold-start performance for deployed ML models.
LinkedIn · Jun 2024
Paul Hetherington said Mystic had achieved record-breaking Docker model-loading speed, including loading AI containers in under 15 seconds, and that the optimization would be open-sourced later.
LinkedIn · Mar 2024
Oscar Rovira announced the Mystic AI BYOC launch on Product Hunt and described it as a serverless experience in the customer’s cloud with minimal DevOps overhead.
Product Hunt · Mar 2024
Product Hunt page for Mystic’s BYOC launch; describes the platform as a high-performance ML deployment product with serverless pay-per-millisecond pricing and BYOC deployment on AWS/Azure/GCP.
Jobs and careers at Mystic
0 jobs at Mystic
No open positions
Check back later for new openings