AI FactoryNeo Cloud

The AI Factory for building your own frontier capabilities.

LegionEdge is an AI/ML lab dedicated to building the frontier models of the future, and to making it easier and more accessible to build your own. Every layer is offered as a service, so the stack you scale on is yours. Most of the work is open.

Start building Explore research

01Protocols
02Data
03Inference
04Training
05Compute

The stack

01Protocols
02Data
03Inference
04Training
05Compute

What we offer

Run on demand-scaled capacity.

The scaling story

From your data to a model you can operate.

Building useful AI takes more than access to a model. It takes the right data, a training method that fits the task, and infrastructure that fits your product.

LegionEdge brings those disciplines together. We help you develop the capability inside your business, with a clear path from the first experiment to deployment.

How we can help

experiment → productionone path, one fleet

01
Data
Curate, generate and validate the corpus the task needs.
Datasets
02
Train
Fine-tune an open model or plan a full training run.
Fine-tuning Model TrainingPreview
03
Evaluate
Measure against a baseline before anything ships.
EvaluationsPreview
04
Serve
Live
Put the model behind an API, shared or dedicated.
Serverless Inference Dedicated Endpoints

Runs on one fleet

GPU Cloud Dedicated ClustersPreview

NVIDIA H200

The workload comes first.
Start with the task, the data, and the constraints. We establish an evaluation baseline before choosing a model or committing to a training run.
The layers work together.
Data quality, training, and serving are connected decisions. Work with one team across the stack, or bring us in for the part you need.
The work stays useful.
The engagement includes the agreed datasets, weights, evaluations, and implementation details. Your team can inspect the result and keep building on it.

Services

Specialist work at every stage.

Engage the lab for one part of your stack or the full project. Each engagement starts with a defined scope, evaluation criteria, and a practical handoff.

Not sure which fits? Describe the problem and the lab will scope it with you.

Engagement brief

01 / 05

Managed datasets

Start with our open corpora or commission data for your domain. Generation, curation, and validation are designed around the task you want to teach.

You bringYour task + domain
We doGenerate. Curate. Validate.
You leave withA corpus built for the task

You receive

A validated corpus
Provenance for every source
A pipeline you can rerun

Scope a dataset How managed datasets works

Engagement brief

02 / 05

Model training

Turn domain knowledge into a model built for your workload. We shape the data, run the training, and measure what changed — with the weights and evaluation tools to keep building.

You bringYour data + domain
We doCurate. Train. Evaluate.
You leave withA specialized model

You receive

A dataset built for the task
Model artifacts and run history
An evaluation you can rerun
A practical deployment handoff

Scope a training run How model training works

Engagement brief

03 / 05

Inference optimization

Make inference fit your product. We tune model size and the serving path around your hardware, traffic, and quality requirements — then show you the tradeoffs against a real baseline.

You bringModel + workload
We doProfile. Optimize. Validate.
You leave withA deployment that fits

You receive

A workload and hardware baseline
An optimized model candidate
A tuned serving configuration
Evidence for the deployment decision

Review your inference stack How inference optimization works

Engagement brief

04 / 05

Protocol design

Give models and tools a shared contract. We turn your interfaces, data, and interaction rules into a precise specification, working code, and tests your team can build against.

You bringSystems + interactions
We doSpecify. Implement. Test.
You leave withA shared contract

You receive

A precise interface specification
A working reference implementation
A conformance test suite
An adoption and versioning plan

Map your interface How protocol design works

Engagement brief

05 / 05

Research partnerships

Bring an open question or a difficult technical decision. Work directly with our researchers to review the evidence, test promising approaches, and turn what you learn into a concrete next step.

You bringYour research question
We doReview. Experiment. Learn.
You leave withEvidence for the next step

You receive

A focused research brief
Technical reviews with clear actions
Experiments your team can revisit
A decision and transfer package

Explore a partnership How research partnerships works

Underneath every engagement

Dedicated compute

GPU capacity for training and serving. Choose the infrastructure your workload needs, with clear hourly pricing.

NVIDIA H200141 GB HBM3e$4.79/hr
Reserve now for Q1 2027New
- NVIDIA GB300288 GB HBM3eQ1 2027
- NVIDIA B300288 GB HBM3eQ1 2027

Starting at

$4.79per GPU-hour

Billed by the hour, on demand or reserved.

Capacity & pricing

Products

Research in practice.

Our work in context protocols and agent systems becomes tools for designing, developing, and shipping software.

All products

A canvas for ideas. Code you can build on.

Move from a written brief to an editable interface. Shape the layout on a visual canvas, work with design tokens, and generate code from the same design.

Multi-agent canvas
Built-in design system
Production-ready code

Explore Nokuva

An editor that works with your context.

A Rust-native IDE and local desktop tools built around our context protocols. Bring project context into the coding workflow, where your AI assistant can use it.

Rust-native and fast
Context protocols built in
Your own models and endpoints

Explore Tavoc

Deployment across clouds.

Application deployment and infrastructure operations, informed by our agent systems research. Bring your product across clouds without making each one a separate workflow.

One spec for every cloud
Autonomous rollbacks
Cost-aware placement

Explore Foltrac

Version control for agent fleets.

A Rust core with copy-on-write snapshots, scoped views, and optimistic merge queues for concurrent work. Open source under GPL 3.0.

Copy-on-write snapshots
Optimistic merge queues
GPL 3.0 core

Explore W0rkTree

Developer documentation

Documentation for the work ahead.

Working examples and technical references for connecting an application, training a model, and putting it into production.

Browse all documentation

Base URL: inference.legionedge.ai/v1
Keys: lek_…
Compatible with: OpenAI SDKs

cURL

Python

TypeScript

POST /v1/chat/completions

1curl https://inference.legionedge.ai/v1/chat/completions \
2  -H "Authorization: Bearer $LEGIONEDGE_API_KEY" \
3  -H "Content-Type: application/json" \
4  -d '{
5    "model": "qwen3-4b-instruct",
6    "messages": [
7      { "role": "user", "content": "Explain LoRA in one paragraph." }
8    ]
9  }'

Keys are issued per workspace from a service account, so apps never act as a person.

Open source

The stack we run on, published in the open.

Identity, compute, data rooms and version control. LegionEdge is built on four projects we maintain as public repositories, so you can read them, install them and run them yourself.

Newsroom

Latest from the lab.

Product releases, research, and engineering notes from the team.

News7

Engineering13

Articles7

AllAll news

Get started

Start with a workload.

Connect to the platform directly, or work with our team on a specific technical challenge.

Self-service

LegionEdge Cloud

Create a workspace, explore the available models, and connect your application. The documentation takes you through your first request.

Turn on only the services you need
Pay per token or per GPU-hour
One account, one bill

Create a workspace Read the docs

Custom engagements

Work with our team

Bring your data requirements, model goals, or infrastructure constraints. We'll define the scope, deliverables, and next steps together.

Scoped with an evaluation baseline
Dedicated clusters and reserved GPUs
A named engineer from day one

Discuss your project sales@legionedge.ai

The AI Factory for building your own frontier capabilities.

Our partners

What we offer

From your data to a model you can operate.

Data

Train

Evaluate

Serve

The workload comes first.

The layers work together.

The work stays useful.

Specialist work at every stage.

Managed datasets

Model training

Inference optimization

Protocol design

Research partnerships

Dedicated compute

Research in practice.

A canvas for ideas. Code you can build on.

An editor that works with your context.

Deployment across clouds.

Version control for agent fleets.

Documentation for the work ahead.

The stack we run on, published in the open.

Latest from the lab.

Start with a workload.

LegionEdge Cloud

Work with our team