Several established programs provide cloud-computing credits to academic researchers at no cost. Each entry below sets out what the program is, who is eligible, how to apply, and how the credit supports work in Mira. Amounts were confirmed in mid-2026; each program’s own page remains authoritative.

Working with Mira Methods

Several of the programs below provide cloud-computing credits that can be used in conjunction with Mira to make it more accessible. A researcher awarded such credits can put them toward the compute needed to run Mira, so that using the application costs nothing; the Google Cloud research grant, in particular, can be applied for in cooperation with the Mira Methods team.

A Google Cloud research grant obtained in cooperation with Mira Methods also helps the wider project. Part of the award is pooled to sustain the shared infrastructure — for example, the compute that serves the open-source models Mira runs — so that the tool remains available to other researchers at no charge.

Most collaboration is disciplinary rather than technical: using Mira in your own research, testing it, and advising on what humanistic and social-scientific work actually needs. That guidance informs the project’s ongoing development, including its work on open-source language models and its multi-agent harness, which the Mira Methods team builds and maintains. Collaborators with the relevant skills and interest are welcome to contribute to that development directly, though it is not expected of everyone.

Google Cloud Research Credits

Google’s academic program provides cloud credits for research projects. It is open to PhD students, postdoctoral researchers, and faculty at accredited institutions in participating countries; eligibility and the list of participating countries are defined by Google, and you can confirm your eligibility on their page. PhD students may apply once a year for up to $1,000; faculty and postdoctoral researchers may apply for up to $5,000. A faculty-led application — for example, in collaboration on Mira’s development — may therefore request the higher amount.

To apply:

1. Create a Google Cloud account at cloud.google.com/free using your university email. This also provides the $300 trial described below.
2. In the Cloud Console, create a billing account (menu → Billing). This is required before applying, and nothing is charged.
3. Write a one-page research proposal describing what you are investigating and how you will use the credits — for example, GPU-hours to serve an open-source research model behind the Mira platform. Google’s pricing calculator will help you estimate the costs the application asks for.
4. Apply through Research Credits at edu.google.com. Applications are considered on a rolling basis, with decisions in roughly six to eight weeks.
5. Once granted, the credits attach to your billing account automatically. Rent a GPU virtual machine (Compute Engine → Create instance → GPUs), or a small application server, and follow the self-hosting guide.

Google Cloud free trial

This is the standard Google Cloud welcome credit. It requires no proposal and no academic affiliation; anyone creating their first Google Cloud account receives it.

1. Go to cloud.google.com/free and choose “Get started for free”.
2. Sign in with any Google account and add a card. This is an identity check only; the trial does not charge automatically, and real billing begins only if you explicitly upgrade.
3. The $300 is available immediately and lasts ninety days. It can be spent as the research credits above are — on a GPU virtual machine for model serving, or a small server hosting your own Mira.

DigitalOcean credit

DigitalOcean is a cloud provider often used for small servers. New accounts receive $200 in credit, valid for sixty days, with one credit per person or payment method.

1. Create an account at digitalocean.com. The $200, sixty-day offer is standard for new sign-ups; it may appear through a referral banner, but the amount is the same.
2. Add a card or PayPal to activate the credit. You are not charged until the credit is exhausted and you continue using the services.
3. Create a “Droplet” — DigitalOcean’s term for a small server. The roughly $24–48 per month sizes run the Mira application comfortably, or a GPU Droplet can serve a model.
4. Follow the self-hosting guide on that machine. The $200 covers about two months of an application server, enough to run a private Mira deployment for evaluation.

Help with an application

If you would rather gain access to Mira by collaborating with us as a researcher-collaborator than pay for beta access yourself, we are glad to help with the application process for these programs — from choosing a suitable one to drafting the research proposal. Write to [email protected], or apply to join as a researcher-collaborator.

Program terms change, and the linked pages are authoritative. None of the links above are affiliate links, and Mira receives nothing from any of these programs.