Skip to content

Releases

v0.3.0

Choose a tag to compare

@raven-clown raven-clown released this 20 Nov 17:45
46ae9eb

What's Changed

Changed

  • Breaking: ${VAR} in the config file fills in part of one value and can no longer add YAML structure. A variable that held a list or a map, such as targets: ${TARGETS}, is now read as a string and the file is refused. Move such lists into the file and keep only secrets in variables (#136). by @raven-clown in #160

Added

  • --config-test flag to validate a configuration file, environment variables, provider routes and settings without starting the server, binding listening ports, or creating a database file (#103). by @nikhil-mewar in #113
  • Every response carries X-Request-ID: the caller's value when it is 1 to 128 printable ASCII characters, or a new 32-character hex id. The same id is recorded with usage and sent to the provider. by @brevitywithorion in #114
  • Every line the JSON logger writes carries "service":"transponder", so logs from the four projects can be told apart in one place. by @brevitywithorion in #115
  • Mistral AI provider (type: mistral) defaulting to https://api.mistral.ai/v1 and sending api_key as a bearer token. Chat and embeddings (mistral-embed) share that base URL. A base_url still overrides that default. No prices are included, so calls are logged as unpriced. by @brevitywithorion in #109
  • NVIDIA NIM provider (type: nvidia) defaulting to https://integrate.api.nvidia.com/v1 and sending api_key as a bearer token. A base_url still overrides that default. A self-hosted NIM is usually http://localhost:8000/v1. No prices are included, so calls are logged as unpriced. by @brevitywithorion in #116
  • Groq provider (type: groq) defaulting to https://api.groq.com/openai/v1 and sending api_key as a bearer token. A base_url still overrides that default. No prices are included, so calls are logged as unpriced. by @brevitywithorion in #108
  • Cerebras provider (type: cerebras) defaulting to https://api.cerebras.ai/v1 and sending api_key as a bearer token. A base_url still overrides that default. No prices are included, so calls are logged as unpriced. by @brevitywithorion in #110
  • SiliconFlow provider (type: siliconflow) defaulting to https://api.siliconflow.com/v1 and sending api_key as a bearer token. A base_url still overrides that default. The China host is https://api.siliconflow.cn/v1. No prices are included, so calls are logged as unpriced. by @brevitywithorion in #117
  • LM Studio provider (type: lmstudio) defaulting to http://localhost:1234/v1. api_key is optional and, when set, is sent as a bearer token. Targets using lmstudio count as free in usage costs. by @brevitywithorion in #118
  • store.postgres (or TRANSPONDER_STORE__POSTGRES) keeps API keys and usage records in Postgres, so several replicas can share them. The schema is migrated at start, and replicas that start together apply each migration once. by @raven-clown in #160
  • Every replica checks the config_version row in the store every 2 seconds and, when it changed, loads and validates the whole runtime config and swaps it in. A config that fails to load leaves the old one serving and is logged once. Calls already running finish on the config they started with. The admin API will raise the version with each write. by @raven-clown in #160
  • With store.postgres, budgets, token allowances, request rates and concurrency caps are counted in Postgres, so they hold across every replica. Two replicas cannot both spend the last of a budget. limits.store: memory keeps per-process counters for a single replica. When Postgres cannot be reached, limits.on_store_error: local counts in each replica with its share of each limit, and deny answers 503 overloaded. by @raven-clown in #160
  • Images in chat: image_url parts on /v1/chat/completions (a URL or a data: URL) and image blocks on /v1/messages are translated to OpenAI, Anthropic and Gemini targets with text and images kept in order. A target that cannot take an image, such as Gemini with an image at a URL, is skipped. A translated inline image over 5 MiB is a 400 image_too_large, and each image counts as 1,000 input tokens for limits until the provider reports usage (#41). by @raven-clown in #160
  • OpenAI Responses API: POST /v1/responses is passed through to targets that speak the OpenAI API with model changed and store set to false, streamed or not, so the Codex CLI and newer OpenAI SDK code work through transponder. Usage, with cached and reasoning tokens, is read from the final body or from the event that ends the stream, so usage headers, usage records, budgets, limits, access per route and fallback work as they do for chat. Anthropic, Gemini and echo targets, and servers without the endpoint, are skipped. Every caller of a route shares the provider account, so nothing is stored there and previous_response_id, conversation and item_reference items are refused with a 400 unsupported: send the whole conversation in input. A caller with llm.max_input_tokens gets a 400 input_not_countable for a stored prompt or an input_file, which the limit could not measure; for other callers a file counts as 1,000 tokens like an image. background: true is refused, since its result and usage could not be fetched through the gateway (#42). by @raven-clown in #160
  • Every response, on the gateway's port and on ops_listen_addr, carries X-Content-Type-Options: nosniff, so a browser never runs an error body that quotes a caller's input as HTML. by @raven-clown in #160
  • Every caller has one of the roles admin, viewer or member, and the role of an API key now comes with the caller on each request, ready for the admin API and access per route. A key stored without a role is a member and calls the gateway as before. by @raven-clown in #160
  • Sign-in with OpenID Connect next to API keys: set auth.oidc.issuer and auth.oidc.audience, and a JWT in Authorization: Bearer or x-api-key is checked for its signature, issuer, audience and expiry. user_claim (default sub), teams_claim (default groups), roles_claim with role_mapping, jwks_url and clock_skew (default 1m) map the token to a user, teams and a role. The issuer's keys are cached for an hour and read again for an unknown key id at most once every 10 seconds. With no keys fetched yet and the issuer down, token callers get 503 auth_unavailable and /readyz reports oidc. When role_mapping names an admin, the first start no longer prints an admin key. by @raven-clown in #160
  • Admin API under /admin/v1/, described by api/openapi.yaml and served at /admin/v1/openapi.yaml: GET /status, API keys (GET, POST, PATCH, DELETE and POST /keys/{name}/rotate, the secret shown once), usage totals grouped by user, team, key, project, model, provider, route or day (GET /usage), single usage records (GET /requests) and the audit log (GET /audit). viewer may read and admin may write. Lists are cursor paginated, errors name the field, and If-Match with the item's version refuses a write over someone else's change with 409 version_conflict. by @raven-clown in #160
  • Providers, routes, budgets and limit profiles can be created, changed and deleted through the admin API (/admin/v1/providers, /routes, /budgets, /profiles), so a new setup needs no YAML. Each write is live on the replica that took it before it answers and on every other replica within 2 seconds. A write that would leave a config that does not load is refused with 400 invalid_config and the old one keeps serving. Deleting a provider a route uses is 409 in_use. by @raven-clown in #160
  • Provider keys sent to the admin API are sealed with AES-256-GCM under secrets_key (TRANSPONDER_SECRETS_KEY) before they are stored, and reads only show whether one is set and its last four characters. With SQLite and no secrets_key, one is generated into transponder.key next to the database; with Postgres it must be set, the same on every replica. by @raven-clown in #160
  • The providers, routes, budgets and profiles of the YAML file are now stored at start and shown by the admin API as managed_by_file, read-only there. Removing one from the file and restarting removes it. by @raven-clown in #160
  • Audit log: every write through the admin API records who made it, what changed (the item before and after, without secrets), when, and the request id, in the same transaction as the write, on SQLite and Postgres. by @raven-clown in #160
  • GET /metrics serves Prometheus metrics: transponder_requests_total, transponder_request_duration_seconds, transponder_inflight_requests, transponder_limited_total and transponder_events_dropped_total, plus transponder_tokens_total and transponder_cost_usd_total by provider and model, transponder_upstream_duration_seconds, transponder_fallbacks_total, transponder_unpriced_requests_total, transponder_config_reloads_total and transponder_build_info. No label carries a user, a key or a request id. On the gateway's port /metrics needs a key or token with the viewer or admin role. by @raven-clown in #160
  • Tracing over OTLP: set otel.endpoint to an OTLP/HTTP collector (port 4318) and each call is exported as one span with a child span for every provider it tried, carrying the provider, model, status, tokens and cost, and never prompt or completion text. otel.headers adds headers for a collector that needs a key, and otel.sample_ratio samples the traces that start at the gateway. A caller's traceparent and tracestate are kept: the call joins the caller's trace and the provider is sent the attempt's span. Off by default. An otel.endpoint that is not an http or https URL is now a config error, where it used to be ignored. by @raven-clown in #160
  • demo/: docker compose up starts transponder in front of a fake provider with calls flowing through it, Prometheus, Jaeger, and a Grafana dashboard of requests, latency, tokens, c...
Read more

v0.2.4

Choose a tag to compare

@raven-clown raven-clown released this 14 Nov 13:47
29327bf

What's Changed

Changed

  • With limits on, a call reserves the price of its route's dearest target instead of the first one, so a free first target cannot let calls through that a paid fallback would bill. by @raven-clown in #106
  • POST /v1/messages/count_tokens goes to the first Anthropic target of the route, wherever it is in the list. by @raven-clown in #106

Added

  • xAI (Grok) provider (type: xai) defaulting to https://api.x.ai/v1 and sending api_key as a bearer token. A base_url still overrides that default. by @brevitywithorion in #101
  • xAI reasoning tokens are counted apart from completion tokens, as in its chat completion usage, and priced at the output rate. by @brevitywithorion in #101
  • Prices for grok-4.7, grok-4.6, grok-4.5 and grok-4.3, using the rate for a prompt under 200k tokens on the global endpoint. Longer prompts, and the US endpoint, cost more than these rates. by @brevitywithorion in #101
  • Vercel AI Gateway provider (type: vercel) defaulting to https://ai-gateway.vercel.sh/v1 and sending api_key as a bearer token. A base_url still overrides that default. No prices are included, so calls are logged as unpriced. by @brevitywithorion in #102
  • Fallback: a route's targets are tried in order, and the next one takes over when a target is unreachable, times out, answers 408, 429, a 5xx or 529, rejects the gateway's key, fails a stream before its first event, or cannot take the request. A request the provider rejects as wrong is not retried elsewhere, and nothing falls back once a byte has reached the caller. by @raven-clown in #106
  • Cooldown: after settings.cooldown_failures (3) failures in a row a target is tried last for settings.cooldown (30s). by @raven-clown in #106
  • X-Transponder-Provider, X-Transponder-Model and X-Transponder-Fallbacks on routed responses, and fallbacks in the usage record. by @raven-clown in #106
  • GET /v1/usage shows the caller what they have used and have left of each budget and token allowance that applies to them. by @raven-clown in #106

Fixed

  • A tool call id made by Gemini is shortened to one OpenAI accepts when a later turn of the conversation goes to an OpenAI target. by @raven-clown in #106

New Contributors

Full Changelog: v0.2.3...v0.2.4

v0.2.3

Choose a tag to compare

@raven-clown raven-clown released this 09 Nov 07:25
9848ca4

What's Changed

Changed

  • X-Usage-Output-Tokens no longer includes reasoning tokens for OpenAI and Gemini; they are in X-Usage-Reasoning-Tokens. by @raven-clown in #97
  • Token estimates count every non-Latin character as one and a half tokens, so Thai, CJK and other scripts are counted on the high side rather than undercounted. It is a heuristic; a model with a small vocabulary can still use more. by @raven-clown in #97

Added

  • llama.cpp server provider (type: llamacpp) defaulting to http://localhost:8080/v1 and forwarding api_key as a bearer token. Targets using llamacpp count as free in usage costs. Note that llama-server listens on :8080 by default, so pass --port or configure listen_addr if running both on the same machine. by @nikhil-mewar in #96
  • USD budgets per user (user:* or an exact user), team and key, daily or monthly. A spent budget answers 402 with the subject, the limit and the reset time: code budget_exceeded for OpenAI-dialect callers, and type billing_error with code budget_exceeded for Anthropic-dialect callers. by @raven-clown in #97
  • Profiles in the shared mothership format, from the config or a shared profiles_file, bound to keys, users and teams, with a default_profile: requests rate and concurrency, llm.tokens rate and daily and monthly allowances, llm.cost_usd, and llm.max_input_tokens, llm.max_output_tokens and llm.max_reasoning_tokens. Calls reserve before they run and settle after, and responses carry RateLimit-Policy and RateLimit headers. by @raven-clown in #97
  • Reasoning tokens are counted apart where OpenAI and Gemini report them, priced at the output rate, stored with the usage and sent in X-Usage-Reasoning-Tokens. by @raven-clown in #97
  • Limit counters are rebuilt from this month's usage at start, so a restart does not reset anyone's spend. by @raven-clown in #97

Fixed

  • Anthropic cache writes kept for an hour are priced at their own rate, twice the input price, instead of the five-minute rate. by @raven-clown in #97

Full Changelog: v0.2.2...v0.2.3

v0.2.2

Choose a tag to compare

@raven-clown raven-clown released this 08 Nov 12:38
65d648a

What's Changed

Changed

  • When a provider of another dialect rejects a translated request as invalid (a 4xx it would give again on a retry), the caller now gets that status and the provider's reason in its own dialect instead of a 502. by @raven-clown in #93

Added

  • GET /healthz answers 200 while the process is up, and GET /readyz answers 503 with the names of the dependencies that fail, such as {"status":"unavailable","failed":["store"]}, so Kubernetes and load balancers can restart the process or hold traffic. Neither needs an API key. by @raven-clown in #89
  • POST /v1/messages, the Anthropic Messages API, so Claude Code and the Anthropic SDKs work with only a base URL change. Requests to Anthropic go through as sent, with only the model changed; streams are copied frame by frame and their usage is recorded. Requests to other providers are translated, tools included. by @raven-clown in #93
  • POST /v1/messages/count_tokens, forwarded to the route's Anthropic target. by @raven-clown in #93
  • Anthropic provider (type: anthropic). It forwards the caller's anthropic-version and anthropic-beta. by @raven-clown in #93
  • Gemini provider (type: gemini) for chat, streams and embeddings through the Gemini API with an API key. by @raven-clown in #93
  • OpenAI chat completions can now reach Anthropic and Gemini, and every translated call carries tools, tool calls and the system prompt. by @raven-clown in #93
  • Prices for current Claude and Gemini models. by @raven-clown in #93

Full Changelog: v0.2.1...v0.2.2

v0.2.1

Choose a tag to compare

@raven-clown raven-clown released this 07 Nov 08:46
a2e50a9

What's Changed

Added

  • transponder --check-update lists the releases newer than the running one, how urgent they are and which changes touch the providers in your config. It only reads the release list and never installs anything. by @raven-clown in #84
  • update_check: true logs one update available line at start and once a day when a newer release is out. Off by default. by @raven-clown in #84
  • Releases come with binaries for Linux, macOS and Windows on amd64 and arm64, a release.json describing each change, and a container image at ghcr.io/raven-clown/transponder. by @raven-clown in #84
  • vLLM provider (type: vllm) defaulting to http://localhost:8000/v1 and forwarding api_key as a bearer token. Targets using vllm count as free in usage costs. by @nikhil-mewar in #85

Fixed

  • transponder --version prints the commit and build time, as listed in 0.1.4. The helper was added then, but the flag still printed only the version. by @raven-clown in #83
  • Boolean config keys can be set from the environment, such as TRANSPONDER_UPDATE_CHECK=true. Before, the value was read as text and the config failed to load. by @raven-clown in #84

Full Changelog: v0.2.0...v0.2.1

v0.2.0

Choose a tag to compare

@raven-clown raven-clown released this 06 Nov 10:36
daa7f27

Added

  • OpenRouter provider (type: openrouter) defaulting to https://openrouter.ai/api/v1.
  • An unknown provider type in the config now lists the known types.
  • GET /v1/models lists the route names a caller can ask for, so clients such as Open WebUI can fill their model picker. Wildcard routes and the models a route sends upstream are left out.
  • settings.attempt_timeout (default 60s) bounds how long a streamed call waits for its first event, and settings.request_timeout (default 10m) bounds every call, streams included. Calls that are not streamed get only the request timeout, since a provider sends their headers once the whole answer is ready. Before, one 10 minute limit covered everything, so a stream that never started held the caller for 10 minutes.

Changed

  • The user column of the usage table is now user_id, since user is a reserved word in Postgres. The migration runs at start and keeps every row; queries on transponder.db that read user need user_id instead.

Fixed

  • A provider stream that ends without [DONE] now ends the caller's stream with an error event and estimated usage, instead of looking complete.
  • A stream line or event over 16 MiB ends the stream instead of growing memory.

v0.1.4

Choose a tag to compare

@raven-clown raven-clown released this 06 Nov 11:18
d297c19

What's Changed

Added

  • LiteLLM Proxy works as a provider. Set type: litellm and transponder talks to http://localhost:4000/v1, or your own base_url, sending api_key as a bearer token. by @nikhil-mewar in #78
  • type: ollama providers now send api_key as a bearer token, so an Ollama behind an auth proxy or a hosted Ollama works instead of failing with 401. With no key, no header is sent. by @Satvik-jain in #75
  • transponder --version shows the commit and build time, for example transponder dev (1a2b3c4, 2026-11-04T10:00:00Z), plus , modified for a dirty tree, so a bug report says exactly which build is running. by @Satvik-jain in #76

Fixed

  • A chat request with a wrong field type now says what to fix, such as max_tokens must be an integer, with param set to the field, instead of a Go error naming internal types. It reads the same on every Go version. by @Satvik-jain in #77

New Contributors

Full Changelog: v0.1.3...v0.1.4

v0.1.3

Choose a tag to compare

@raven-clown raven-clown released this 05 Nov 12:31
f560cbb

Thanks to @jay23606 for #72.

v0.1.2

Choose a tag to compare

@raven-clown raven-clown released this 05 Nov 05:34
531c4f8

v0.1.1

Choose a tag to compare

@raven-clown raven-clown released this 05 Nov 04:46
19c5a8b

Fixed

  • Preserve every data line when forwarding multiline streaming events.

Thanks to @snehaapratap for the fix in #28.