Built for real workloads
Move from quick prompts to coding sessions, agent loops and automated workflows with an endpoint designed for sustained use.
Managed access · Custom capacity · Private deployments
Put powerful AI
inside your workflow.
LLM infrastructure for coding, agents and production workloads.Connect your existing tools through an OpenAI-compatible API.
● LIVEOPENAI-COMPATIBLE
base_url https://api.hommmai.net/v1
model kimi-k2.7-code
01IDE & coding tools
Implement · debug · refactor
02Agent workflows
Plan · execute · iterate
03Your applications
Build · integrate · ship
→ One endpoint, from local development to production workflows.
//
Designed for developers building with AI every day. Use your tools. Keep your workflow.
Infrastructure for applied AI
From first prompt to real workflow.
Connect a capable model to the places where work actually happens: coding environments, autonomous agents, internal automation and the products your team is building.
A focused model, a familiar API and a deployment path that can evolve with your requirements.
One familiar API
Connect through an OpenAI-compatible endpoint and keep working with the tools, SDKs and workflows your team already knows.
Deployment that can grow
Start with managed access, then talk to us about custom capacity, private environments and requirements specific to your organisation.
The model
One model.
Chosen for real work.
The launch service starts with one carefully selected model instead of a catalogue that shifts the evaluation work onto your team.
It is intended for coding, reasoning, agentic workflows and general-purpose development tasks.
More models are planned. We will expand the catalogue deliberately as new capabilities prove useful in real workflows.
MODEL / 01● LAUNCH CONFIG
Kimi K2.7 Codemodel: "kimi-k2.7-code"
interface: "openai_compatible"
launch: "q4_2026"
roadmap: "more_models"
Start focused. Expand the catalogue when it adds meaningful capability.
Built for the work
One endpoint. Many ways to use it.
01
Coding agents
Give agents the room to inspect, implement, test and iterate across longer development tasks.
02
IDE assistants
Bring model access into supported editors for implementation, debugging, refactoring and explanation.
03
Automation
Power scripts, internal tools and repeatable agent workflows from the same compatible endpoint.
04
Product integrations
Add model capabilities to internal or customer-facing applications with a familiar API surface.
A service that can evolve
Start managed.
Go further when you need to.
CapabilityWhat you getWhy it matters
AccessOpenAI-compatible APIUse existing clients and SDKs
WorkflowsCoding, agents and automationOne endpoint across your stack
OperationsManaged serviceWe run the shared infrastructure
Commercial modelMonthly subscriptionNo per-token invoicing
Extended needsCustom and private optionsScoped with your team
AI plan configurator
Configure your plan.
We’ll recommend the right model and minimum compatible plan. You can leave the model choice to us.
Early access
Reserve now.
Pay when access starts.
Create your account and finish a secure no-charge checkout through Paddle. Nothing is charged today.
Create your account
Sign in and reserve your place in the launch queue.
Complete Paddle Checkout
Securely add the billing and payment details requested by Paddle. Nothing is charged today.
Get activated
When capacity becomes available, your subscription can be activated at your reserved launch price.
Managed access
Choose the capability
your work needs.
Max uses flat monthly pricing, includes one concurrent active session and is subject to the documented fair-use policy.
Max is the only active offer. Core, Pro, Pro+ and Ultra are currently disabled.
CoreDisabled
Not currently offered
Unavailable
Models
Meta Muse Glimmer 30BDisabled
NVIDIA Nemotron 3.5 LightningDisabled
- Flat monthly pricing
- 1 concurrent active session
- Fast everyday workloads
ProDisabled
Not currently offered
Unavailable
Models
Mistral Small 4Disabled
Ling 3.0 FlashDisabled
- Flat monthly pricing
- 1 concurrent active session
- Tools and documents
Pro+Disabled
Not currently offered
Unavailable
Models
DeepSeek V4 FlashDisabled
- Flat monthly pricing
- 1 concurrent active session
- Advanced reasoning
MaxPrelaunch
Best for Developers
$150/ month
Models
Kimi K2.5Coming soon
Kimi K2.7 CodePrelaunch
- Flat monthly pricing
- 1 concurrent active session
- Advanced coding and agents
Exact model can be updated before launch.
UltraDisabled
Not currently offered
Unavailable
Models
Qwen 2.4T-classDisabled
Kimi K3Disabled
- Flat monthly pricing
- 1 concurrent active session
- Frontier capability
PADDLE / MERCHANT OF RECORDNo charge when you reserve.
Paddle securely handles payment processing, applicable taxes and billing documents. Paid billing starts only when your account is activated.
Custom & private deploy
Infrastructure shaped around your workload.
For teams with specific capacity, isolation or integration requirements, we can scope an environment beyond the standard managed plan.
- 01Dedicated capacity
Resources planned around your workloads, concurrency and service expectations.
- 02Private environments
Deployment and access boundaries designed with your security and data requirements in mind.
- 03Custom integration
Technical discovery for networking, authentication, observability and application integration.
Progressive rollout in Q4 2026
Deployments begin at 500 reservations.
As soon as 500 founding reservations are confirmed, we begin deploying the infrastructure for initial access.
Activations then open in controlled batches as capacity expands with real demand.
DEPLOYMENT THRESHOLD 2H REFRESH
— / 500
Why reserve before launch?
Join the launch
Be among the first teams to connect their development workflows when deployments begin.
No payment today
Add your payment method in Paddle without starting paid billing.
Help shape the service
Early demand helps us prioritise capacity, integrations and the workflows that matter most.
Questions, answered
What to know before you start.
Details about the model, supported workflows, launch access, activation and billing.
What am I paying today?+
Nothing. Completing Paddle Checkout reserves your place and launch offer. Paddle creates a trialing subscription, but paid billing does not begin during the reservation process.
Will my payment method be charged when I reserve?+
No. Paddle securely collects your payment details for future subscription activation. No subscription payment is collected when you reserve.
When will I be charged?+
Your first payment is due when your subscription is activated. If additional card authentication is required, you may be asked to confirm the payment before access is enabled.
When do deployments begin?+
Initial deployments begin as soon as 500 founding reservations are confirmed. Access then opens progressively in controlled batches as infrastructure capacity expands.
How many sessions are included?+
The launch plan includes one active session at a time with access to Kimi K2.7 Code. Additional sessions are not included in this initial plan.
How does fair use work?+
The flat monthly plan is designed for sustained individual development workflows. Proportionate technical controls may apply to protect service quality and shared capacity; any material limits will be documented before activation.
Which model is included?+
The initial launch plan includes Kimi K2.7 Code. We are starting with one model to keep quality and the developer experience focused, with more carefully selected models planned after launch.
Can I use it for coding agents?+
Yes. The service is being designed specifically with coding assistants, agents and developer workflows in mind.
Is this another AI chat app?+
No. The focus is LLM access for your development tools and workflows rather than another standalone chat interface.
Can I cancel my reservation?+
Yes. You can cancel your reservation before activation from your account.
Can I cancel after activation?+
After activation, you can cancel at any time to stop future renewals. Payments already made are non-refundable except where required by law, charged incorrectly or more than once, or where the paid service was not supplied or materially failed to match its description.
Why do you need payment details if you are not charging me?+
A payment method helps distinguish genuine launch demand from a simple email waitlist and makes activation easier when capacity becomes available. Paddle handles the payment details securely.
Why is billing information requested?+
Paddle requests the customer and billing information required to prepare the subscription, calculate applicable taxes and issue billing documents. Full payment credentials are handled only by Paddle.
Who handles checkout and billing?+
Paddle is our Merchant of Record. Paddle handles payment processing, applicable taxes, billing documents and subscription payments. We do not store your full payment details.
Is my payment information stored by you?+
Full payment details are handled by Paddle. The application stores only the identifiers and limited payment-method metadata required to manage your reservation and future subscription.
Jobs
Help build infrastructure for applied AI.
We are building a team around reliable model serving, developer tooling and practical AI systems. We value people who can turn ambitious technical ideas into products that work in the real world.
GENERAL INTERESTREMOTE · INTERNATIONAL
Tell us what you want to build.
We welcome introductions from engineers, infrastructure specialists and product-minded operators. Send us a short note about your experience, the problems you care about and links to relevant work.
Contact info@hommm.net Please do not include sensitive personal information in your first message.Launching in Q4 2026
Bring the model into your workflow.
Reserve managed access for launch, or tell us about the infrastructure your team needs.
Reserve managed access →Request a custom quote