Alex b5296df8a9 feat(pricing): cached-input rates for gpt-5.4-mini and gpt-5.4-nano ...

Checked against OpenAI's pricing page: gpt-5.5 at $5 / $30 (cached $0.50)
was already right. The mini and nano models declared no cached rate, so
cached prompt tokens were billed at the full input rate.

2026-09-21 14:30:48 +01:00

..

agents

fix(quotas): address review: resume claims, agent bucket rule, UI races

2026-09-21 12:44:26 +01:00

alembic

feat(quotas): quota_policies table and a per-call cost on token_usage

2026-09-21 12:11:11 +01:00

api

fix(quotas): address review: resume claims, agent bucket rule, UI races

2026-09-21 12:44:26 +01:00

core

feat(pricing): cached-input rates for gpt-5.4-mini and gpt-5.4-nano

2026-09-21 14:30:48 +01:00

deploy

fix: keep credentials and 5000-digit paths out of the Redis URL errors

2026-09-17 13:34:59 +01:00

devices

refactor: read settings directly instead of getattr with a second default

2026-09-17 11:14:34 +01:00

events

fix(api): address review on the event-loop streams

2026-09-14 10:19:59 +01:00

graphrag

fix(graphrag): drop entities whose name normalizes to nothing

2026-09-20 10:45:05 +01:00

guardrails

refactor: read settings directly instead of getattr with a second default

2026-09-17 11:14:34 +01:00

llm

feat(quotas): quota_policies table and a per-call cost on token_usage

2026-09-21 12:11:11 +01:00

parser

refactor: read settings directly instead of getattr with a second default

2026-09-17 11:14:34 +01:00

prompts

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

quotas

fix(quotas): list unpriced models by recorded cost; integer token limits in status

2026-09-21 12:14:14 +01:00

retriever

fix(retriever): search a graph source classically when its graph answers nothing

2026-09-20 10:45:03 +01:00

sandbox

refactor: read settings directly instead of getattr with a second default

2026-09-17 11:14:34 +01:00

scripts

refactor: read settings directly instead of getattr with a second default

2026-09-17 11:14:34 +01:00

security

chore(deps): anthropic 1.5, openai 3.13, and the move to httpx2

2026-09-12 17:44:18 +01:00

seed

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

services

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

storage

fix(quotas): address review: resume claims, agent bucket rule, UI races

2026-09-21 12:44:26 +01:00

streaming

fix: stability improvements on event loop

2026-09-14 12:52:15 +01:00

stt

feat: air-gapped deployment guide, no implicit downloads

2026-09-15 17:54:24 +01:00

templates

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

tts

feat: air-gapped deployment guide, no implicit downloads

2026-09-15 17:54:24 +01:00

updates

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

vectorstore

fix(worker): know you are in a worker from any thread, not only the task's

2026-09-19 14:07:42 +01:00

__init__.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

__main__.py

fix: run the module when the docsgpt command is not on PATH

2026-09-16 22:45:54 +01:00

alembic.ini

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

app.py

feat(pat): enforce scopes and resource restrictions, deny by default

2026-09-19 23:39:58 +01:00

asgi.py

feat(api): serve long-lived streams on the event loop

2026-09-14 08:33:46 +01:00

auth.py

feat(pat): authenticate personal access tokens

2026-09-19 23:39:58 +01:00

cache.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

celery_init.py

fix(worker): record worker state at startup, for every pool

2026-09-19 14:33:14 +01:00

celeryconfig.py

fix(rename): address review on the package rename

2026-09-07 12:02:07 +01:00

cli.py

feat: a development loop in one command

2026-09-17 11:32:50 +01:00

Dockerfile

feat: serve the UI from the backend image, one-port standalone stack

2026-09-15 22:41:53 +01:00

error.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

gunicorn_conf.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

gunicorn_worker.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

index.faiss

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

index.pkl

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

logging.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

mcp_server.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

pricing.py

feat(pricing): per-million model rates and a cost module

2026-09-21 11:38:22 +01:00

requirements-docling.txt

chore(deps): final upgrade pass (setuptools 84)

2026-09-12 18:21:33 +01:00

requirements-milvus.txt

chore(deps): torch 2.14, transformers 5.17, tokenizers 0.23 (docling extra)

2026-09-12 18:15:03 +01:00

requirements.txt

chore(deps): torch 2.14, transformers 5.17, tokenizers 0.23 (docling extra)

2026-09-12 18:15:03 +01:00

ui.py

feat(package): ship the web UI in the wheel and serve it from the API

2026-09-09 11:09:09 +01:00

upload_limits.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00

usage.py

feat(quotas): quota_policies table and a per-call cost on token_usage

2026-09-21 12:11:11 +01:00

utils.py

refactor: read settings directly instead of getattr with a second default

2026-09-17 11:14:34 +01:00

version.py

chore: 0.21.0

2026-09-16 14:23:26 +01:00

worker.py

feat(quotas): enforce user quotas on chat, agent, scheduled and webhook runs

2026-09-21 12:11:11 +01:00

wsgi.py

refactor: rename the application package to docsgpt

2026-09-07 10:20:43 +01:00