routstr-core

clients

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

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

__init__.py

feat(upstream): add native DeepSeek provider, retire V4 pricing shim

2026-09-30 13:35:01 +00:00

anthropic.py

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

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

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

chore: merge main

2026-10-01 22:57:42 +05:30

cache_breakpoints.py

explicit cache

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

cooldown.py

fix: guard meaningful stream events and scope cooldown failures

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

count_tokens.py

refactor: share x-cashu usage estimation and log full refunds

2026-09-08 22:05:07 +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

refactor: pass outcome contexts instead of a record flag and reuse the shared helpers

2026-10-01 11:21:53 +05:30

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

feat(upstream): add native DeepSeek provider, retire V4 pricing shim

2026-09-30 13:35:01 +00:00

groq.py

revert: remove unrelated repository formatting

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

helpers.py

feat(upstream): add native DeepSeek provider, retire V4 pricing shim

2026-09-30 13:35:01 +00:00

http_client.py

perf: reuse process ssl context for x-cashu clients

2026-09-27 00:26:15 +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

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

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

ollama.py

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

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

openai.py

fix: send max_completion_tokens to OpenAI reasoning models up front

2026-09-30 03:00:16 +02:00

openrouter.py

fix: report openrouter:unknown when OpenRouter omits the serving provider

2026-09-29 09:48:56 +00:00

perplexity.py

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

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

ppqai.py

fix: address review findings on cooldowns, melt sizing, and stream finalization

2026-08-26 23:15:40 +02:00

pricing_resolver.py

fix(pricing): a rate spelled as a boolean is not a rate

2026-08-26 15:44:29 +02:00

rate_limit.py

revert: remove unrelated repository formatting

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

reasoning_effort.py

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

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

request_correction.py

fix: rename params upstreams reject with a named replacement

2026-09-30 02:27:58 +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

terminal_outcome_tracking.py

refactor: drop stream finish tracking the ledger no longer reads

2026-10-01 14:40:14 +05:30

tinfoil_trailer.py

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

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

tinfoil.py

fix: report upstream 5xx as 424 UPSTREAM_UNAVAILABLE instead of node-down

2026-09-25 02:01:13 +02:00

typesafe.py

refactor(typesafe): simplify catalog assembly, seed versioned ids, log fetch failures, drop session report doc

2026-09-20 15:57:55 +02:00

venice.py

fix: make Claude Code work on Venice with prompt caching and gpt-6

2026-09-30 00:13:07 +02:00

xai.py

revert: remove unrelated repository formatting

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