Provider Billing

Available provider playbooks

What each playbook must contain

Every provider file follows the same shape so they're comparable and composable:

1. Requirements — CLI tools, auth, env vars, permissions
2. Known identifiers — account IDs, subscription IDs, billing profiles, regions (real values, not placeholders)
3. Query endpoints — one section per endpoint (realtime usage, billed transactions, invoices, quotas, credit balance)
4. Field semantics — what each response field means, especially ambiguous ones
5. Credit / discount handling — how to tell what's covered by credits, MACC, sponsorship, etc.
6. Deployment operations — how to create/list/inspect/delete resources (if the provider hosts compute/models)
7. Gotchas — every wrong-param-name / permission-denied / regional-naming quirk we've hit
8. Cheat sheet — "Question → query" table at the end
9. Known unknowns — open questions and blockers (permission grants needed, unresolved provisioning errors, etc.)

Template for a new provider

When you add a new provider playbook, copy this structure:

# <Provider> billing via CLI

Validated: <YYYY-MM-DD>. Re-validate if a command returns unexpected results.

## Requirements
- CLI: `<cli-name>` installed and authenticated
- Python3 for JSON wrangling
- Any required environment variables or credentials

## Known identifiers (our account)
<Account / subscription / project / org IDs with real values>

## Querying spend and usage

### 1. <Endpoint name> — <latency / granularity>
<exact curl or CLI command with our identifiers pre-filled>
<sample output fields with semantics>

### 2. <Next endpoint>
...

## Credit / discount handling
<How credits apply, which fields show them, permission requirements>

## Deployment operations (if applicable)
<Create / list / inspect / delete commands>

## Gotchas
- <Specific error → root cause → fix>
- <...>

## Question → query cheat sheet
| Question | Endpoint | Example |

## Known unknowns
- <Blocker 1>
- <Blocker 2>

Adding a new provider playbook

1. Create providers/<name>.md using the template above
2. Add the row to the table in this file
3. Fill in only what you validated in this session — don't guess at endpoints
4. If something is unverified, mark it ⚠️ UNVERIFIED inline so the next session knows to check it

Related skills

• spending-analysis — Pollinations revenue side (Polar orders + Tinybird usage). Pair with this skill when answering "how much did we spend vs. how much did we earn."
• tinybird-deploy — Tinybird deployment mechanics. This skill will eventually host the cost side of Tinybird.
• enter-services — enter.pollinations.ai deployment. Related for EC2 and Cloudflare Worker spend questions.