Skip to content

Chore/token usage#21

Open
samwilliamsxc wants to merge 2 commits into
mainfrom
chore/token-usage
Open

Chore/token usage#21
samwilliamsxc wants to merge 2 commits into
mainfrom
chore/token-usage

Conversation

@samwilliamsxc

Copy link
Copy Markdown
Contributor

No description provided.

samwilliamsxc and others added 2 commits July 24, 2026 13:17
Per-call cost is now usage-based (scales with data read/written), so a single
fixed per-endpoint cost no longer applies. Rename the OpenAPI extension
x-token-cost → x-average-token-cost, set values to TBD pending real averages,
and relabel the generated "Token cost" column → "Average token cost".

Rewrite rate-limits.md for the new model: 100k default bucket, advisory
metering (depleted bucket returns a token_warning, not a 429), server_busy,
monthly per-license pool, extended-access unthrottled. Reconcile the same facts
in openapi.yaml (meta schema + error enum), responses-and-errors, restricted,
and the guides/examples that hardcoded the old fixed-cost model.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Merge the whoami endpoint from main and populate the previously-TBD
x-average-token-cost values from the measured per-endpoint averages
(avg_total_kb, used as-is since 1 token ≈ 1 KB). Welcome = 0 (uncharged),
single-anchor GET and whoami = 1. Regenerated all reference tables + README.

Conflicts resolved: openapi.yaml/getting-started.md kept both sides (whoami
additions + the average-cost rework); README.md and reference/users.md
regenerated from the merged spec.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@samwilliamsxc

Copy link
Copy Markdown
Contributor Author

Staging this for release on 7/27

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant