9qeklajcandGitHub 87d12fc4e2 Merge pull request #820 from Routstr/feat/decision-model-billing ...

feat: decision model billing and X-Cashu provider URLs

2026-10-20 17:43:33 +02:00

..

clients

fix: switch openai NOT_GIVEN to omit for openai 2.x compat

2026-04-26 22:42:39 +02:00

__init__.py

merge: sync main into venice image pricing, rebase image_pricing migration onto cashu swaps head

2026-10-08 11:33:09 +02:00

anthropic.py

fix(messages): preserve output_config.effort and validate it consistently

2026-10-19 08:15:34 +00:00

auto_topup.py

chore: trim redundant comments

2026-08-26 22:51:09 +02:00

azure.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00

base.py

Merge pull request #820 from Routstr/feat/decision-model-billing

2026-10-20 17:43:33 +02:00

cache_breakpoints.py

explicit cache

2026-06-13 23:40:39 +02:00

certification_cache.py

fix: harden provider certification and native pricing

2026-10-03 20:00:26 +02:00

certification_probe.py

fix: harden provider certification and native pricing

2026-10-03 20:00:26 +02:00

certification.py

fix: use PPQ's published cache rates and keep pricing rows ok without model rows

2026-10-06 17:27:19 +02:00

cooldown.py

fix: guard meaningful stream events and scope cooldown failures

2026-09-29 02:58:40 +02:00

count_tokens.py

fix: bill decision 200s only with answers and count embedding token ids

2026-10-17 12:45:30 +02:00

deepseek.py

fix: add backoff so litellm's deepseek /v1/messages stream works on prod installs

2026-10-01 11:49:59 +02:00

ehbp.py

fix(ehbp): relay encrypted enclave error responses verbatim

2026-10-17 17:18:20 +00:00

fireworks.py

refactor(upstream): resolve a provider's own row by primary key

2026-07-01 12:53:31 +02:00

gemini_messages.py

perf: reduce request latency

2026-09-26 02:41:19 +02:00

gemini.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00

generic.py

fix: send Anthropic block-list system as a string to generic upstreams

2026-10-12 22:06:56 +02:00

groq.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00

helpers.py

fix: harden provider certification and native pricing

2026-10-03 20:00:26 +02:00

http_client.py

perf: reuse process ssl context for x-cashu clients

2026-09-27 00:26:15 +02:00

image_catalog.py

fix: route OpenRouter images to /images, bill cheapest variant without cost, refuse paid Venice options

2026-10-12 00:16:12 +02:00

image_generation.py

fix: reserve OpenAI default quality at the dearest step and bill oversize or costless image responses

2026-10-09 17:28:24 +02:00

json_codec.py

perf: use orjson in streaming hot path and cache provider url

2026-09-26 15:51:52 +02:00

litellm_routing.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00

messages_dispatch.py

test: cover mid-stream litellm failures on buffered messages paths

2026-10-01 11:49:57 +02:00

model_paths.py

fix: harden provider certification and native pricing

2026-10-03 20:00:26 +02:00

ollama.py

feat: expose per-model reasoning effort on /v1/models

2026-09-09 13:31:19 +02:00

openai.py

merge: sync main into venice image pricing, rebase image_pricing migration onto cashu swaps head

2026-10-08 11:33:09 +02:00

openrouter.py

fix: route OpenRouter images to /images, bill cheapest variant without cost, refuse paid Venice options

2026-10-12 00:16:12 +02:00

perplexity.py

feat(perplexity): add integration attribution

2026-09-20 14:22:11 -05:00

ppqai.py

fix(ppqai): stop exposing private/ enclave routes

2026-10-19 09:23:37 +00:00

pricing_resolver.py

fix: close image billing review gaps

2026-10-08 15:17:57 +02:00

rate_limit.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00

reasoning_effort.py

fix(messages): only count output_config as reasoning when it has an effort

2026-10-19 12:30:55 +00:00

request_correction.py

fix: never strip request identity or nested params, read top-level detail

2026-10-11 18:01:49 +02:00

routstr.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00

sse_splitter.py

perf: split streaming SSE events in linear time

2026-09-26 16:02:11 +02:00

stream_ownership.py

fix: bound billed request lifetimes and recover abandoned reservations

2026-09-30 20:01:56 +08:00

stream_timeout.py

fix: guard meaningful stream events and scope cooldown failures

2026-09-29 02:58:40 +02:00

tinfoil_trailer.py

fix(ehbp): classify upstream TLS/connect failures as upstream-scoped

2026-09-27 17:44:38 +05:30

tinfoil.py

fix(tinfoil): publish the effort map alone and stop reshaping valid efforts

2026-10-19 07:54:37 +00:00

together.py

add together ai cache price

2026-10-14 15:00:39 +02:00

typesafe.py

feat: bill decision models without usage and confine them to systemone

2026-10-06 16:49:17 +02:00

venice.py

fix: route OpenRouter images to /images, bill cheapest variant without cost, refuse paid Venice options

2026-10-12 00:16:12 +02:00

xai.py

fix: price upscales by factor, reserve Venice variants, drop unpriced xAI image models

2026-10-08 12:12:01 +02:00