Applications reviewed in waves

Build with open-source models.
Apply for inference grants.

Leantoken is preparing inference grants for AI builders who need reliable access to open-source models. Selected teams receive at least $1,000 in inference credits through leantoken.tech to build, test, launch, and scale AI products with less infrastructure pressure.

Learn how it works
Grant fit

Receive a minimum $1,000 grant in inference credits.

Spend the grant within 30 days against the metrics declared in your six-pager.

Use the grant through code agents, agent swarms, your own app, our API, or any workflow that helps you reach your stated goal.

Mature teams can request larger grants, and strong projects can earn follow-on grants.

Inference should not stop good AI products from reaching users

Many AI teams can build strong prototypes, but production inference costs make it hard to test ideas properly, serve users, and keep iterating.

The Leantoken Inference Grant Program is designed for builders who are ready to ship. We provide selected projects with inference access so they can focus on product progress, user feedback, and measurable traction.

Use the grant your way

Spend the grant on the path that gets you to the goal

The grant is not limited to one prescribed workflow. Selected teams can use the allocation in the way that best supports the goals they declare in their application.

We care about the outcome. The important question is whether you can turn the grant into the goals you declare in your six-pager.

Run code agents such as OpenCode, Claude Code, Codex, or other tools.

Run swarms of agents such as OpenClaw, Hermes, Mirofish, or other orchestration systems.

Use the Leantoken API inside your own product, internal tool, research workflow, or customer-facing app.

Use any available service or workflow that helps you make measurable product progress.

Who it is for

Builders with a serious plan and a clear reason to use inference

Early-stage startups

Pre-seed and seed teams building AI-native products, agents, automation, vertical SaaS, developer tools, research workflows, or model-powered consumer apps.

Indie developers and solo founders

Independent builders with a prototype, public demo, active repository, waitlist, or a specific feature ready to launch.

Existing products

Teams already running AI features who can describe current usage, infrastructure constraints, and the growth they want to support.

AI studios and agencies

Teams shipping AI systems repeatedly and looking for reliable inference infrastructure across multiple deployments.

Researchers and labs

Research groups, independent researchers, and technical teams that need inference for experiments, demos, benchmarks, or public tools.

Open-source maintainers

Maintainers of AI frameworks, libraries, templates, agents, and developer tools that help builders adopt open-source models.

How the grant works

Apply, spend the allocation, show progress, then earn larger support

1

Apply

Tell us what you are building, who it is for, how you will use open-source model inference, and what you expect to ship during the 30-day grant period.

2

Get reviewed

Applications are reviewed in waves for clear use cases, credible execution, and projects where inference access can make a visible difference.

3

Receive inference access

Selected projects receive a Leantoken grant allocation. The minimum grant is $1,000 in inference credits.

4

Show progress

During the 30-day grant period, teams should spend the full grant allocation and show shipped features, active users, usage growth, technical learnings, or other traction.

5

Earn larger support

Projects that complete their obligations after using the initial $1,000 grant can be considered for a larger follow-on grant.

6

Continue with Leantoken

Projects that use the grant well may receive larger grants, extended limits, discounted access, technical support, or a custom infrastructure plan.

What we look for

Strong applications are specific, concrete, and focused on execution

Clear use case

Explain what you are building and why open-source model inference matters to the product.

Strong execution

Show that you can ship quickly, communicate clearly, and turn infrastructure access into product progress.

Path to users

You do not need to be large, but you should have a realistic plan for reaching a specific audience.

Measurable progress

Define what success looks like during the 30-day grant period: users, shipped features, API usage, demos, research output, or revenue signals.

Long-term fit

We are especially interested in projects that may continue using Leantoken after the grant.

What you can use the grant for

  • Building and testing AI product features.
  • Running open-source model inference for prototypes.
  • Running code agents such as OpenCode, Claude Code, Codex, or similar tools.
  • Running swarms of agents through orchestration systems such as OpenClaw, Hermes, Mirofish, or similar tools.
  • Using the Leantoken API inside your own app, service, internal tool, or research workflow.
  • Supporting early users during launch.
  • Comparing model behavior across real product tasks.
  • Expanding an existing AI workflow.
  • Powering demos, benchmarks, or research tools.
  • Integrating inference into open-source projects.
Application

What to include in your application

Project overview

Describe the product, tool, research project, or open-source work you are building.

Target users

Tell us who the project is for and what problem it solves.

Current stage

Share whether the project is an idea, prototype, private beta, public launch, revenue-generating product, research project, or open-source tool.

Inference needs

Explain how you plan to use open-source model inference and what kind of usage you expect.

Grant plan

Tell us what you will build, test, launch, or improve during the 30-day grant period. The initial grant is $1,000 in inference credits, so explain how you will use the full allocation within 30 days. If you are requesting a larger initial grant, state the amount, explain why it is needed, and describe the outcomes it will produce.

Success metrics

Define how progress should be measured: users, shipped features, API usage, feedback, repository activity, benchmarks, research output, or revenue signals.

Six-pager

Applications should include a six-pager in the Amazon style. You can use sixpagermemo.com as a practical guide. The six-pager is a major part of the decision because it explains the project, team, grant usage plan, and expected outcomes. Teams that need more than the initial $1,000 grant should explain the requested amount, expected usage, and 30-day outcomes in the six-pager.

01

What are you building?

02

Who is it for?

03

What stage is the project at?

04

Do you have a demo, repository, users, customers, or public materials?

05

How will you use open-source model inference?

06

How much inference do you expect to use during the grant?

07

What is blocking or slowing you down today?

08

What will you ship or prove during the 30-day grant period?

09

Which metrics should show that the grant was useful after 30 days?

10

Can you use the full initial $1,000 grant within 30 days?

11

Are you requesting a larger initial grant? If yes, what amount and why?

12

Would you consider continuing with Leantoken after the grant?

13

Would you consider a deeper infrastructure partnership if the project shows strong traction?

Before you apply

Document handling and NDA policy

Document handling

We respect the work behind every application and will treat submitted documents carefully. Application materials will not be shared with third parties.

NDA policy

We will not sign NDA agreements for grant applications. If an NDA is required before you can share the project, please do not apply.

Built for teams that are ready to ship

Leantoken grants are designed for practical builders. The best applications are specific, concrete, and focused on execution. You do not need a polished company. You do need a serious plan.

If access to inference would help you launch faster, serve users, or prove that your AI product works, we want to hear from you.

FAQ

Common questions

What does the grant provide?

Selected projects receive inference access through Leantoken for open-source models. The minimum grant is $1,000 in inference credits.

How long do I have to use the grant?

Each project has 30 days to spend the full grant allocation and achieve the metrics declared in its six-pager.

Can I receive more than $1,000?

Yes. Mature teams can request a larger initial grant in their six-pager. Projects that complete their obligations after using the initial $1,000 grant can also be considered for a larger follow-on grant.

Who can apply?

Startups, solo founders, independent developers, researchers, AI studios, agencies, and open-source maintainers can apply.

Do I need to have users already?

No. Existing users help, but they are not required. We also consider strong prototypes, technical demos, research projects, and open-source tools.

What makes an application strong?

A strong application explains the product clearly, describes how inference will be used, defines what will be shipped or measured during the 30-day grant period, and includes a convincing six-pager.

Do I need to submit a six-pager?

Yes. Applications should include a six-pager in the Amazon style. You can use sixpagermemo.com as a guide.

How can I use the grant?

You can use it in the way that best supports your stated goals: code agents, agent swarms, your own application, Leantoken API usage, research workflows, demos, benchmarks, or other services available to you.

Can existing products apply?

Yes. If you already have users or production usage, explain your current workflow and how Leantoken can help you support growth.

What happens after the grant?

Projects that show strong progress may continue with Leantoken through larger follow-on grants, extended access, discounted usage, technical support, or a custom plan.

Is this only for companies?

No. Individuals, research groups, open-source projects, and studios can apply.

Will you sign an NDA?

No. We will treat application materials carefully and will not share them with third parties, but we will not sign NDA agreements for grant applications.

Apply for a Leantoken Inference Grant

If you are building with open-source models and inference access would help you move faster, apply for the grant program. Tell us what you are building, what you need, and what you will prove with the grant. Include a six-pager that makes the case clearly.