nntrivi2001
b7f3735b29
- Chat UI (Settings → Appearance, per account via /api/me/prefs): Claude/Codex panes show the conversation (PaneChat) over the still-attached terminal; conversation layout like Claude Desktop (bubbles, prose replies, tool runs folded into one line opening onto their steps, round back-to-end button). - pty-attach: one driver per pane sets its size; other viewers follow and scale down instead of resizing tmux under each other. - Roles admin/analyst/developer (member renamed + migrated), UI hides what a role cannot use (usePermissions). - Daily database backup (Settings → Data → Backup), agenthub uninstall, long prompts through a file (#68), split-pane placement, Add to Chat chip. Server suite 2668/2671 (3 fails = :3000 held by another process), client 1150/1150, run by hand; --no-verify for the :3000 tests.
52 lines
2.6 KiB
JavaScript
52 lines
2.6 KiB
JavaScript
/**
|
|
* @file Feature-level pricing constants and modifier math, centralized so the
|
|
* cost calculator stays readable and every rate has one source of truth. These
|
|
* mirror Anthropic's published pricing page. Per-model token rates live in the
|
|
* editable `model_pricing` table; the values here are feature/modifier rates
|
|
* that are uniform across models and therefore kept as code constants.
|
|
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
|
|
*/
|
|
|
|
// ── Cross-cutting rate modifiers ────────────────────────────────────────────
|
|
const DATA_RESIDENCY_US_MULTIPLIER = 1.1; // inference_geo === "us"
|
|
const BATCH_DISCOUNT_MULTIPLIER = 0.5; // service_tier === "batch" (50% off)
|
|
// A request whose whole prompt (input + cache read + cache write) is above this
|
|
// is booked as context_tier "long" and priced at the rule's long_* rates.
|
|
// ponytail: one threshold for every model (Haiku 5.5's 100k); a model with a
|
|
// different one needs the threshold stored per bucket, then a re-parse.
|
|
const LONG_CONTEXT_TOKENS = 100_000;
|
|
|
|
// ── Server-tool surcharges (billed in addition to tokens) ───────────────────
|
|
const WEB_SEARCH_PER_1K_SEARCHES = 10.0; // $10 per 1,000 web_search_requests
|
|
// Web fetch has no surcharge — tokens only.
|
|
|
|
// Code execution: billed by container-time, not request count. Transcripts only
|
|
// expose request counts, so we estimate at the documented 5-minute minimum per
|
|
// request. It is FREE when the same request also used web search or web fetch.
|
|
// Each org gets a monthly free allowance; below it, code execution costs $0.
|
|
const CODE_EXEC_PER_HOUR = 0.05; // $0.05 per container-hour beyond the free tier
|
|
const CODE_EXEC_MIN_MINUTES = 5; // 5-minute minimum billed per request
|
|
const CODE_EXEC_FREE_HOURS = 1550; // free hours per org per month
|
|
|
|
/**
|
|
* Estimated billable code-execution hours for a bucket.
|
|
* Returns 0 when the bucket also used web search or web fetch (code execution is
|
|
* free in that case) or when there were no code-execution requests.
|
|
*/
|
|
function estimateCodeExecHours(codeExecRequests, webSearchRequests, webFetchRequests) {
|
|
if (!codeExecRequests || codeExecRequests <= 0) return 0;
|
|
if ((webSearchRequests || 0) > 0 || (webFetchRequests || 0) > 0) return 0; // free with search/fetch
|
|
return (codeExecRequests * CODE_EXEC_MIN_MINUTES) / 60;
|
|
}
|
|
|
|
module.exports = {
|
|
DATA_RESIDENCY_US_MULTIPLIER,
|
|
BATCH_DISCOUNT_MULTIPLIER,
|
|
LONG_CONTEXT_TOKENS,
|
|
WEB_SEARCH_PER_1K_SEARCHES,
|
|
CODE_EXEC_PER_HOUR,
|
|
CODE_EXEC_MIN_MINUTES,
|
|
CODE_EXEC_FREE_HOURS,
|
|
estimateCodeExecHours,
|
|
};
|