SpareTokens

Company and lab guide

A small pilot with a capped project key and a security review. We help you from the first email to your first verified result.

Talk to usTeam overview

Time
A few days for a small pilot, about 30 minutes of hands-on work per person
Cost
You pay only the API spend you cap. There is no price list yet, and we agree anything else in writing before it applies.
You need
One owner for the pilot, a project or service-account key with a spend limit, and one to five machines.
You get
A verified result from your own team, a record of what ran and what it cost, and a written summary at the end.

Where things stand. SpareTokens is a private pilot. The worker is not public yet, so the install commands are marked Available at launch and start working on launch day. Everything else on this page is how the worker behaves today. The waitlist is how you get an invite.

Before you start

This is the hands-on version of the team overview. It follows a small pilot from first email to first verified result. Team pages, a team leaderboard, organization controls and an automated impact report are planned and not built yet, so the pilot is deliberately small and run by hand with you.

Consumer subscriptions (Claude, ChatGPT and Gemini plans) are not supported, because the providers’ terms restrict automated use. The worker refuses a subscription login and points you here. Use an API key or a local model.

Step 1: Tell us who you are and what you want to try

About 10 minutes

Send a short email. A few lines are enough: your organization, who owns the pilot, the providers you use, and the monthly spend you are comfortable capping.

Email hello@sparetokens.ai

We reply with a plain plan, the questions your security reviewer is likely to ask, and the security summary, so the review can start straight away.

Step 2: Agree the scope and the caps

About 1 day

Together we pick a small scope: one to five machines or keys, a monthly cap for each, the task types in scope and any quiet hours. A cap is enforced on each machine before each task, so it cannot be exceeded by accident. Write the limits down. They go into each person’s giving rule in step 4.

If something goes wrong
  • Your provider has team limits: set the same limit at the provider. A cap in the worker and a cap at the provider protect you twice.
  • You need approval for automated API use: read your provider’s terms and your own policy first. Subscriptions are not supported; API keys and local models are.

Step 3: Create a dedicated project key

About 15 minutes

In your provider’s console, create a project or workspace just for SpareTokens, set its spend limit, and create a key inside it. Share it with the person who sets up each machine through your normal secret manager, never in chat or email. See project and service-account keys.

The worker reads the key by name from the machine’s environment. It is never written into SpareTokens config, and it never leaves the machine.

Step 4: Set up the first machine

About 20 minutes

One person follows the API credit guide on the first machine, with your cap. In short, once the worker is installed:

pipx install sparetokens

Available at launch The install command works from launch day, and pilot participants get access before then.

sparetokens setup
sparetokens give pilot --provider anthropic-api --fixed 20 --period month --key-ref env:ANTHROPIC_API_KEY
sparetokens preview pilot

Replace the amount with your agreed cap. Each person appears under a username by default. Real names are shown only if that person opts in.

A terminal running sparetokens give and sparetokens preview, which prints given, expired and kept amounts for an example week.
Choose a rule, then preview it. The numbers are an example week, not a forecast.
If something goes wrong
  • Anything the worker prints is explained on troubleshooting.
  • Run sparetokens doctor and paste its output into your email to us. It is read-only and never contains a key.

Step 5: Let your security reviewer look

About 1 hour

Reviewers can read exactly what a task will do before it runs, and check the public log themselves:

sparetokens inspect TASK_ID
sparetokens verify-log

The security summary links to the trust page, the security model and the list of known limits. Questions go to hello@sparetokens.ai.

Step 6: Run one task and roll out

About 5 minutes

sparetokens run --budget-usd 2 --max-tasks 1 --allow-bash

You will see

SpareTokens worker: provider local (local), tier 1, capabilities ['bash', 'solver', 'lean'], budget n/a, max tasks 1
No giving rule for local (it is not a configured source), so only these limits apply: --max-tasks 1.
Task 1 [code_fill] Write a small function: abundant count (instance 0) (cap n/a)
  submitted as result 2 (cost $0.0000, tokens 2000, spend flag: none)
  verification: verified: all 8 answers recomputed exactly
Done. Tasks submitted: 1. Dollars: $0.0000, tokens: 2000, wall: 0s.

Real output from the real worker, captured against a throwaway coordinator with throwaway signing keys. Paths are shortened to ~. This capture used a scripted test model, so the cost reads $0.0000. With your key, the same lines show what the task cost.

When that works, repeat steps 3 and 4 for the other machines, or pause the pilot with sparetokens pause 2h while you decide.

If something goes wrong
  • A task was refused: the worker checks signatures, the pack release and the public log first. A refusal is the safety working, and we want to hear about it.
  • You hit a cap: that is by design. Raise it in your giving rule when you are ready.

Your first verified result

Your history shows the verified line, and your pilot summary will list it:

sparetokens history --limit 5

You will see

2026-11-21 06:34  local          task 1  code_fill        verified   $0.0000  2000 tokens  fake-codefill
2026-11-21 06:34  local-ollama   task 13  canary           verified   $0.0000  1000 tokens  fake-codefill

Real output from the real worker, captured against a throwaway coordinator with throwaway signing keys. Paths are shortened to ~.

At the end of the pilot we send a short written summary: tasks run, results verified, spend as reported by your workers, and anything that went wrong. We will not publish your name or results without asking you first.

What “verified” means. A machine check written and owned by SpareTokens passed. It does not by itself mean a result is new or important, and no discovery has been claimed. See what has been verified so far and how each check works.

Be told when the worker is ready

Create an account and you are in line for the private beta. We email you when it is your turn.

Get started