Have your agent do some good.
Your Claude plan already includes a free, dedicated agent for headless work — and most of it sits idle. Lend it to nonprofits: real tasks, run by your own machine. No new spend, no busywork, just capacity pointed at work that matters.
Put your idle agent to work.
Your Claude plan includes a free, dedicated Agent SDK credit built for headless claude -p work — separate from your normal usage and mostly left on the table. Point that included allowance at vetted tasks, let your local runner do the work, and keep a verifiable record of everything you gave.
- Runs on your own machine & subscription
- You set the budget; you stay in control
- A running tally of every contribution
Up and running in three commands:
# sign in with GitHub (opens your browser) npx github:Barneyjm/givework.dev login # set how much of your own credit to donate (¢/mo) npx github:Barneyjm/givework.dev budget set 2000 # do donated work on your own claude -p EXECUTOR=claude npx github:Barneyjm/givework.dev run --watch
Needs the claude CLI installed & logged in — that session is the donated capacity.
Tell us what you need. In plain words.
Email what you're drowning in — summaries, data cleanup, drafts, categorization. We turn it into right-sized tasks and route them to volunteers. You never pick a model, write a prompt, or see a bill.
- Email hello@givework.dev — that's the whole start
- No cost, no software to install
- Real, reviewable results back to you
Two sides, one tally.
Say it plainly
Send a plain-language email describing the work. No technical terms required — write it the way you'd ask a colleague.
We break it down
Your request is decomposed into small, well-scoped tasks, each sized so a single volunteer can finish it cleanly.
Volunteers do it
Developers' AI agents pick up the tasks and complete them. You get reviewable results — and never a charge.
Sign in & run one command
npx github:Barneyjm/givework.dev login, set a budget, then run. Your machine, your subscription — work is checked out and done locally; nothing leaves your control.
Check out a task
Each task reserves against a budget you set, so a run can never overspend. Submit the result when it's done.
Keep the record
Each checkout and submission adds to your running tally — so you can always see your own impact and point to the work you've done.
Clear books, both ways.
Givework keeps a simple running tally of what's been donated — so you can see your own impact, stay inside the budget you set, and nonprofits know the work adds up. Plain bookkeeping, nothing hidden.
| checkout · reserve | +500 |
| submit · actual | −380 |
| accept · signed off | 0 |
| donated this task | 120¢ |
The actual pipeline.
Intake
A nonprofit we work with emails a plain-language need. We partner with established nonprofits we know — not anonymous submitters — so there's a real organization behind every request. No account, no form, no model names.
Decompose
A small, free model running on our side breaks the request into structured tasks — each with a prompt, an output schema, acceptance criteria, a model, and a hard cost cap in cents.
Review & publish
A human approves the draft before anything runs. Approved tasks enter an open pool; nothing executes on raw, unreviewed input.
Reserve
A volunteer's runner authenticates with a scoped token, claims a task, and reserves its cap against a budget. A row-level lock with a database invariant (reserved + spent ≤ budget) means no run can ever overspend.
Execute on idle capacity
The runner does the work with claude -p headless on the volunteer's own machine. Inputs never leave their control; there is no API key anywhere in the system.
Submit & reconcile
Actual cost — taken from the CLI's own metering — replaces the reservation. Checkout, submit, and accept are each tracked, then the reviewed result goes back to the nonprofit.
# a decomposed, reviewed task { "title": "Summarize intake form", "spec": { "prompt": "…", "output_schema": { "summary": "string" }, "acceptance": "one paragraph, plain English" }, "model": "claude-opus-4-8", "max_cost_cents": 500 } # run on the volunteer's own machine $ claude -p --output-format json \ --model claude-opus-4-8 \ --json-schema schema.json < prompt
The donated capacity isn't usage anyone bought, and it isn't a loophole. Every Pro, Max, Team, and Enterprise plan now includes a free monthly Claude Agent SDK credit — a separate allowance, on top of your normal usage limits, meant for exactly this kind of headless claude -p work. Anthropic supports it explicitly; it's a benefit you're given, not capacity they're counting on you to waste. Givework simply points that included, mostly-unused credit at nonprofit work: no new spend, no API keys, all on the volunteer's machine. The article steers shared production automation toward a paid API key — which is exactly why this stays personal: every volunteer donates their own individual credit from their own logged-in session, never a shared or pooled key.