v1.6.4

​

✨ Features

• Sarvam AI Provider - Added Sarvam AI as a first-class provider with chat, text-to-speech, and speech-to-text support (thanks @Purvi09!)
• ElevenLabs Sound Effects - Added text-to-sound generation support via /v1/sound-generation (thanks @SecretSun!)
• Bedrock Project Scoping - Added optional project_id to Bedrock and Bedrock Mantle key configs with per-alias overrides for Bedrock, Bedrock Mantle, and Vertex, plus UI support
• Trace Redaction - Phase-scoped redaction and revealing, transient redaction data field for guardrails, and trace content redaction before connector export
• Durable Background Jobs - New sidekiq background-job table, store methods, and runner with recovery and reaper; cost recalculation migrated to a durable, resumable job with polling instead of SSE
• Audit Log Object Storage - S3/GCS object storage config schema for audit log archival
• Alerting Configuration - Alerting schema in config.schema.json with declarative channels and CEL-based rules, Helm chart support, and enterprise fallback pages
• Model Catalog Pricing - Added pricing data to the model catalog (thanks @johnbrett!)
• Canonical Model Names - Dashboard model rankings now show canonical model names instead of inference-profile IDs (thanks @satyamkrishna!)
• OAuth2 Hardening - Allowlist for private-use redirect URI schemes (RFC 8252 §7.1) and a shouldSweep gate on the OAuth2 sweep worker
• Mirrored Schema Support - schema_url / BIFROST_SCHEMA_URL for mirrored schema locations in isolated deployments
• Vertex Single-Region Config - Enforce single-region configuration in Vertex key config
• Helm Chart Updates - bifrost.alerting, audit-log object storage, postgresql.external.port string support, and bifrost.mcp.toolGroups[*].id

​

🐞 Fixed

• Governance Rate-Limit Reset CPU - Guards against invalid reset timeouts, parallelized resting-budget flows only when absolutely required, and fixed the calendar-based alignment qualifier
• Masked Key Persistence - Never persist masked provider key previews to config storage (thanks @eyeveil!)
• OpenShift Arbitrary UIDs - Build-time group-0 ownership with no runtime chown (thanks @eyeveil!)
• Passthrough Virtual Key Attribution - Passthrough calls via the Azure api-key header now attribute to the virtual key (thanks @eyeveil!)
• Rerank for Custom Providers - /v1/rerank now works with custom OpenAI-compatible providers (thanks @eyeveil!)
• Responses Stream Usage - Persist stream usage when providers omit or reuse sequence numbers (thanks @eyeveil!)
• Wildcard allowed_models Repair - Repair bare wildcard allowed_models rows that broke admin provider updates (thanks @eyeveil!)
• Streaming Error Panic - Nil-safe tracing span lookup prevents panics on streaming errors (thanks @eyeveil!)
• Anthropic Tool ID Sanitization - Sanitize tool_use/tool_result ids to Anthropic’s charset (thanks @Shaik-Sirajuddin!)
• Realtime Transcription Sessions - Support GA transcription-type sessions in POST /v1/realtime/client_secrets (thanks @Shaik-Sirajuddin!)
• Diarized Transcription - Support diarized_json segments and ElevenLabs speaker passthrough (thanks @Shaik-Sirajuddin!)
• Model Discovery - Skip disabled keys when scheduling model-discovery fetches (thanks @Shaik-Sirajuddin!)
• MCP Timeout Placeholder - Show the real global default in the MCP tool execution timeout placeholder (thanks @Shaik-Sirajuddin!)
• Redacted Thinking Round-Trip - Round-trip Anthropic redacted_thinking blocks on the Responses surface (thanks @fus3r!)
• Streaming Accumulation - Preserve citation annotations and finish_reason in the accumulated streaming response (thanks @fus3r!)
• Gemini Grounded Streaming - Reset web-search flag when recycling pooled stream state so web_search_call items keep emitting (thanks @fus3r!)
• Vertex gs:// Images - Pass through gs:// image URLs on Vertex Gemini
• Bedrock Truncation Signal - Signal max_output_tokens truncation on the Responses API (thanks @jeremym-tanium!)
• Bedrock Reasoning Config - Preserve reasoning_config on cross-provider translation so fallbacks keep extended thinking (thanks @Purvi09!)
• Anthropic tool_search - Forward and rebuild server-side tool_search on the Responses path (thanks @ws4charlie!)
• OpenAI Responses Input - Strip role from non-message input items (thanks @nettee!) and serialize compaction request input correctly (thanks @mcclurmc!)
• additional_tools Support - Added additional_tools message type support, preserving nested tool types on /v1/responses
• Plugin Stream Errors - Emit structured plugin stream errors on integration routes (thanks @jeffhos!)
• Pooled Object Hygiene - Zero pooled ChannelMessage references on release and sweep orphaned deferred spans in trace store TTL cleanup (thanks @citrocat!)
• Hybrid Log Token Usage - Rebuild token usage from denormalized columns in hybrid log list (thanks @G-XD!)
• MCP Tool Ordering - Deterministic MCP tool ordering for prompt cache stability
• MCP Inline-Auth Links - Warn callers not to truncate the #t= temp-token fragment (thanks @MarcusPeng!)
• Gemini Fixes - Web search options map to Google Search grounding, file upload MIME types preserved, and video reference fields map to instances (thanks @vojthor!)
• OpenAI Parameters - Honor service tier in chat completion and cap max reasoning effort
• Anthropic Costing - Correct inference geo cost and cache rate for fast mode
• SecretVar Parsing - Parse SecretVar JSON with ref/env_var fields even when value is absent
• Telemetry - Forward request id and trace id, reduce metrics cardinality explosion risk, and send status codes on OTEL metrics
• Dashboard - Preserve active time period when applying dimension filters, adjust bucket size thresholds for month-range durations, show user popover with preferred_username fallback, and filter provider-level keys from the prompt manager selector (thanks @rlex!)
• API Key Provider Selection - Fixed provider selection for API keys
• Azure Auth Headers - Pass Azure auth headers in helpers
• Stream Delta Schema - Added ExtraContent to ChatStreamResponseChoiceDelta (thanks @nghodkicisco!)

​

🐙 Closed GitHub Issues

• #2347 - MCP tool ordering is non-deterministic, breaking prefix-based prompt caching
• #3455 - Segfault/nil dereference panic in Bedrock provider
• #4318 - allowed_models persisted as bare ”*” string blocks subsequent provider updates
• #4353 - config.db corruption from masked-key preview in provider_configs JSON column
• #4367 - Image incompatible with OpenShift arbitrary UIDs
• #4402 - Vertex provider drops image blocks whose URL uses gs:// scheme
• #4477 - Passthrough calls using a Virtual Key log as actual key
• #4679 - Bedrock Responses API does not signal max_output_tokens truncation
• #4689 - Custom providers cannot set budget
• #4712 - ElevenLabs sound effects (/v1/sound-generation)
• #4780 - Anthropic server-side tool_search results are dropped on /v1/responses
• #4834 - /v1/rerank is not available with custom providers
• #4846 - Responses stream usage present in response.completed but not persisted in LLM Logs
• #4851 - Governance rate-limit reset causes high CPU in BumpRateLimitUsage
• #4870 - Pooled ChannelMessage retains request body, context, and undelivered response while idle
• #4940 - Show canonical model names instead of Bedrock inference-profile IDs in Model Rankings
• #4963 - Streaming finish_reason dropped from the accumulated (logged) response
• #5002 - gpt-4o-transcribe-diarize transcription fails due to string segment IDs
• #5013 - OpenAI /responses/compact input serialized as a JSON object causing 400
• #5027 - MCP Tool Execution Timeout placeholder shows 0 instead of real global default
• #5036 - Plugin StreamInterceptionError is flattened on integration routes
• #5037 - Disabled keys break provider model discovery
• #5051 - Add Sarvam AI provider (chat + TTS/STT)
• #5061 - Streaming responses drop citation annotations from the accumulated message
• #5093 - Streaming /v1/responses drops Anthropic redacted_thinking blocks
• #5097 - Anthropic rejects replayed tool_use/tool_result ids from non-conforming upstream providers
• #5100 - additional_tools loses nested tool types on /v1/responses
• #5101 - Chat-to-Responses tool replay sends role on function_call input items
• #5108 - Bedrock reasoning_config silently dropped on cross-provider translation
• #5113 - Gemini/Vertex streaming stops emitting web_search_call items after first grounded request

• feat: add Sarvam AI provider with chat, text-to-speech, and speech-to-text support (thanks @Purvi09!)
• feat: add ElevenLabs sound effects (text-to-sound) support (thanks @SecretSun!)
• feat: add ProjectID to Bedrock and Bedrock Mantle key configs with per-alias overrides for Bedrock, Bedrock Mantle, and Vertex
• feat: add additional_tools message type support
• feat: force single-region config in Vertex key config
• feat: phase-scoped redaction and revealing with transient redaction data field, plus ClearPausedStreamBuffer for pause-accumulate stream flows
• fix: map web search options to Google Search grounding in the Gemini API
• fix: parse SecretVar JSON with ref/env_var fields even when value is absent
• fix: cap max reasoning effort in OpenAI
• fix: honor service tier in OpenAI chat completion
• fix: support GA transcription-type sessions in POST /v1/realtime/client_secrets (thanks @Shaik-Sirajuddin!)
• fix: support diarized_json transcription segments and ElevenLabs speaker passthrough (thanks @Shaik-Sirajuddin!)
• fix: sanitize tool_use/tool_result ids to Anthropic’s charset (thanks @Shaik-Sirajuddin!)
• fix: never persist masked provider key previews (thanks @eyeveil!)
• fix: make tracing span lookup nil-safe to prevent panic on streaming errors (thanks @eyeveil!)
• fix: enable rerank for custom OpenAI-compatible providers (thanks @eyeveil!)
• fix: forward and rebuild Anthropic server-side tool_search on the Responses path (thanks @ws4charlie!)
• fix: zero pooled ChannelMessage references on release to avoid pinning request bodies (thanks @citrocat!)
• fix: round-trip Anthropic redacted_thinking blocks on the responses surface so multi-turn tool use with redacted reasoning can be replayed (thanks @fus3r!)
• fix: reset HasEmittedWebSearch when recycling pooled Gemini responses stream state so grounded streaming requests keep emitting web_search_call items (thanks @fus3r!)
• fix: omit role from OpenAI Responses non-message items (thanks @nettee!)
• fix: serialize OpenAI compaction request input correctly (thanks @mcclurmc!)
• fix: preserve reasoning_config on Bedrock cross-provider translation (thanks @Purvi09!)
• fix: preserve Gemini file upload MIME types for GenAI file URI completions
• fix: Gemini video reference fields map to instances (thanks @vojthor!)
• fix: accept object-valued tool-call arguments (e.g. tool_search_call) on the Responses API streaming path
• fix: recover from idle-timeout timer-goroutine panic that could crash the process
• fix: deterministic MCP tool ordering for prompt cache stability
• fix: pass through gs:// image URLs on Vertex Gemini
• fix: signal Bedrock max_output_tokens truncation on Responses API (thanks @jeremym-tanium!)
• fix: warn callers not to truncate the #t= temp-token fragment on MCP inline-auth links (thanks @MarcusPeng!)
• fix: add ExtraContent to ChatStreamResponseChoiceDelta (thanks @nghodkicisco!)
• fix: correct inference geo cost on Anthropic and cache rate for fast mode
• fix: send MIME type when specified
• fix: send status code on OTEL metrics
• fix: reduce telemetry metrics cardinality explosion risk
• fix: race conditions in tracer span locks
• fix: pass container block from Anthropic API
• fix: pass Azure auth headers in helpers
• chore: upgrade ClickHouse client library

• feat: add durable background-job sidekiq table, store methods, and runner with recovery and reaper
• feat: pass created-by user ID and runner ID through sidekiq job lifecycle, add GetInFlightSidekiqJobByKind to config store interface
• feat: show canonical model names in dashboard model rankings (thanks @satyamkrishna!)
• feat: redact trace content before connector export with transient redaction data field
• feat: force single-region config in Vertex key config
• fix: persist Responses stream usage when providers omit or reuse sequence numbers (thanks @eyeveil!)
• fix: fold streamed output_text.annotation.added events into the accumulated responses message so citations survive in logging, observability, and cache (thanks @fus3r!)
• fix: keep the streaming finish_reason in the accumulated response when a provider forwards it on a content chunk (thanks @fus3r!)
• fix: sweep orphaned deferred spans in trace store TTL cleanup (thanks @citrocat!)
• fix: rebuild token usage from denormalized columns in hybrid log list (thanks @G-XD!)
• fix: repair bare wildcard allowed_models rows that break admin provider updates (thanks @eyeveil!)
• fix: use AutoMigrate and add runner_id/created_by_user_id columns to sidekiq table migration
• fix: match model filter on canonical_model_name and restore routing info for cost recalculation
• fix: forward ScopedDB from HybridLogStore
• fix: race conditions in tracer span locks
• chore: upgrade ClickHouse client library

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• fix: parallelize resting-budget flows only when absolutely required to reduce lock contention
• fix: add guards against invalid rate-limit reset timeouts
• fix: calendar-based alignment qualifier for budget periods
• fix: attribute passthrough virtual keys sent via Azure api-key header (thanks @eyeveil!)
• chore: upgraded core to v1.6.4 and framework to v1.4.4

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• feat: migrate cost recalculation to durable background sidekiq job with resume and dedup
• feat: redact trace content before connector export
• fix: match model filter on canonical_model_name and restore routing info for cost recalculation
• chore: upgraded core to v1.6.4 and framework to v1.4.4

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• fix: send status code on OTEL metrics
• chore: upgraded core to v1.6.4 and framework to v1.4.4

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• chore: upgraded core to v1.6.4 and framework to v1.4.4

• feat: add missing connector attributes
• fix: reduce metrics cardinality explosion risk
• fix: forward request id and trace id through telemetry
• chore: upgraded core to v1.6.4 and framework to v1.4.4