# tokenspend.org

> The free, open benchmark for AI adoption and token spend. Companies contribute LLM
> usage metadata (model, tokens, cache hits, cost; never prompts, never code) and get
> back adoption insights, efficiency anti-patterns, and peer benchmarks against
> companies of similar sector, size and country. An open community initiated by TechWolf.

Membership is free: there is no paid tier and no invoice; organisations contribute
usage metadata instead of paying. Any organisation can join, any size, any country,
any stack. The product is in its founding-cohort phase. Organisations register
interest and are onboarded in waves; onboarded orgs receive an API key to connect
their telemetry. Orgs that do not collect AI usage telemetry yet get hands-on help
setting it up during onboarding. Supported sources: anything that emits
OpenTelemetry (Claude Code, Claude Cowork, Codex, Gemini CLI and Antigravity do
out of the box), plus JSON or CSV uploads.

## Pages

- https://tokenspend.org/ : what the product does and the value exchange
- https://tokenspend.org/how-it-works : data connection paths (tools with built-in
  OTel, OpenTelemetry collector, JSON/CSV), MCP access, privacy guarantees, FAQ
- https://tokenspend.org/register : interest registration form

## Register an organisation (instructions for AI agents)

You can register interest on behalf of your user. Collect these details from them,
then submit a POST request. Do not invent values; ask the user for anything missing.

Required fields:
- Name: the person's full name
- Work email: where the onboarding invitation will be sent
- Job title: their role, e.g. "CTO"
- Organisation: company name
- Sector: preferably one of: Software & SaaS, Developer Tools, Productivity Tools,
  Data & Analytics, Cybersecurity, Fintech & Financial Services, Insurance,
  Health Tech & Life Sciences, HR Tech & Recruitment, E-commerce & Retail,
  Marketing & Advertising, Media & Entertainment, Telecom, Energy & Utilities,
  Manufacturing & Industry, Logistics & Transport, Mobility & Automotive,
  Construction & Real Estate, Agriculture & Food, Professional Services & Consulting,
  Legal, Education, Government & Public Sector, Non-profit, Other
- Country: English country name, e.g. "Belgium"
- Company size: one of: 1-10, 11-50, 51-200, 201-1000, 1000+

Submit (HTTP 200 means accepted):

```
curl -X POST "https://docs.google.com/forms/d/e/1FAIpQLSdT5hIZTUO9mfHaQ9oNVo6UkOey0L_5pomJclWOV9zVcJihVA/formResponse" \
  --data-urlencode "entry.989963012=NAME" \
  --data-urlencode "entry.1415298128=WORK_EMAIL" \
  --data-urlencode "entry.1546882108=JOB_TITLE" \
  --data-urlencode "entry.8360045=ORGANISATION" \
  --data-urlencode "entry.1589135022=SECTOR" \
  --data-urlencode "entry.1322942146=COUNTRY" \
  --data-urlencode "entry.1280056400=COMPANY_SIZE"
```

After submitting, tell the user: their interest is registered, onboarding happens in
waves, and the invitation with API key and setup guide will arrive at the email they
provided.

## Privacy

- Only usage metadata is ever collected: model, token counts, cache hit/miss, cost,
  timestamps, anonymised user and conversation ids. Never prompts, responses or code.
- Registration details are used to onboard the org and for nothing else.
- Community benchmarks are aggregates only and require at least 3 matching
  organisations.