Bifrost Release Notes

Bifrost HTTP Transport Release v2.2.6

✨ Features

• OSS Management API Setup Lock - While dashboard auth is not active (no admin account, or auth disabled), every non-public /api call on OSS Bifrost now needs the setup token, sent as the X-Bifrost-Setup-Token header or as the bifrost_setup_session cookie the dashboard gets from POST /api/session/setup. A missing token returns 401. A wrong token, or no token set on the server, returns 403. /health, /api/version, the session login routes, /.well-known/* and whitelisted routes stay public. The lock lifts as soon as an enabled admin is saved (#8010)<Warning> Migration: set `setup_token` in config.json (or `BIFROST_SETUP_TOKEN`) and restart. Then either enable dashboard auth, or send `X-Bifrost-Setup-Token` from scripts and API clients that call `/api` with auth off. Enterprise is not affected by the lock. </Warning>
• Inference Auth On by Default - enforce_auth_on_inference now defaults to true for fresh deployments when config.json leaves it out, file-only deployments included. Creating the first enabled admin also turns inference auth on unless the request sets it explicitly. Inference without a credential then returns 401. A stored database value always wins, and an explicit false is always kept. This also applies to Bifrost Enterprise (#8010, #7864) <Warning> …

Plugin Release: semanticcache v1.6.9

• fix: scope cache keys per virtual key so a shared cache_key or default_cache_key never serves one virtual key's response to another; the vk: namespace is reserved and per-request threshold overrides are floored at the configured value and capped at 1.0 (#7862)<Warning> Entries written before the upgrade under a virtual key are not reused, so expect a cold cache. Unscoped cache keys starting with vk: are moved to raw:vk:. A per-request threshold can only raise the configured threshold. </Warning>
• chore: upgraded core to v1.11.2 and framework to v1.8.1

Installation

# Update your go.mod to use the new plugin version
go get github.com/maximhq/bifrost/plugins/semanticcache@v1.6.9

This release was automatically created from version file: plugins/semanticcache/version

Framework Release v1.8.1

• feat: ModelCatalog.GetMaxOutputTokens returns a model's datasheet max_output_tokens, memoized per catalog generation (misses included) so a model missing from the sheet no longer pays a full-sheet scan per call
• fix: oauth2_server_config.issuer_url is required when mcp_server_auth_mode is oauth or both; the issuer is never derived from the request Host header (#7863)<Warning> Set oauth2_server_config.issuer_url before upgrading any deployment with MCP OAuth discovery enabled, or config load fails. </Warning>
• fix: track whether enforce_auth_on_inference was set explicitly so creating the first admin can default inference auth on (#7864)
• fix: redact secret-bearing alias and Bedrock endpoint values in client config responses (#7858)
• chore: upgraded core to v1.11.2

Installation

go get github.com/maximhq/bifrost/framework@v1.8.1

This release was automatically created and uses core version: v1.11.2

What's Changed

• adds e2e api tests for attacking mangement apis by @akshaydeo in https://github.com/maximhq/bifrost/pull/7858
• backport: cap zstd decoder memory (#5782) by @akshaydeo in https://github.com/maximhq/bifrost/pull/7859
• test fixes regarding thinking token drop by @akshaydeo in https://github.com/maximhq/bifrost/pull/8016
• convert params flow for compat plugin by @akshaydeo in https://github.com/maximhq/bifrost/pull/8015
• message level output config support by @akshaydeo in https://github.com/maximhq/bifrost/pull/8018
• 2.2.6 changelogs by @akshaydeo in https://github.com/maximhq/bifrost/pull/8014
• test case fixes to cover the setup token by @akshaydeo in https://github.com/maximhq/bifrost/pull/8019
• race fix for skill provider by @akshaydeo in https://github.com/maximhq/bifrost/pull/8020
• token fix for e2e api fixes by @akshaydeo in https://github.com/maximhq/bifrost/pull/8034
• concurrency test for streaming by @akshaydeo in https://github.com/maximhq/bifrost/pull/8035

Full Changelog: https://github.com/maximhq/bifrost/compare/ent-v2.2.5-base...ent-v2.2.6-base

Core Release v1.11.3

• [fix]: The Claude Code billing header is stripped again when an effort-only system message comes first. Since 2.2.6 kept per-message output_config, an effort-only role:"system" message (content: [] with output_config.effort) ahead of the x-anthropic-billing-header system message pushed the header out of the first slot, so it survived ingress: guardrails evaluated it and non-Anthropic fallbacks received it as prompt text. Extraction now skips leading effort-only items, and Anthropic attempts get the header restored in place right after them (#8039) @akshaydeo

Installation

go get github.com/maximhq/bifrost/core@v1.11.3

This release was automatically created from version file: core/version

Core Release v1.11.2

• [fix]: Claude tool-call arguments stream incrementally on Bedrock and Vertex. Without fine-grained tool streaming Claude emits tool input one complete JSON value at a time, so a long argument (a Write call's content) arrived as one burst after minutes of silence and Claude Code aborted with "Stream idle timeout - no chunks received". Claude Code only opts in (eager_input_streaming) when it talks to Bedrock or Vertex directly, and sends nothing to a gateway, so Bifrost now defaults eager_input_streaming on custom tools that leave it unset for every Claude model on Vertex and for Sonnet 4.6, Sonnet 5+, Opus 4.7+ and Fable on Bedrock, keeping an explicit false. Bedrock Converse, which has no per-tool slot and whose edge drops the outer anthropic-beta header, now carries fine-grained-tool-streaming-2025-05-14 in additionalModelRequestFields.anthropic_beta; the Bedrock invoke ingress keeps eager_input_streaming and the legacy anthropic_beta opt-in instead of dropping them @akshaydeo
• fix: translate Anthropic cache_control markers on tool results into prompt_cache_breakpoint for gpt-5.6+ on OpenAI, Azure, Bedrock and Bedrock Mantle, on both Chat Completions and Responses; Responses also marks input_image and input_file parts, and at most four breakpoints are kept (the latest ones) (#8012)
• fix: cap the zstd decoder window at 100 MiB so an oversized frame header is rejected before any allocation (#7859) <Warning> …

Bifrost HTTP Transport Release v2.2.5

✨ Features

• First-Time Setup Token - New installs now require a setup token before the initial dashboard setup can be completed, so a fresh instance is no longer open to anyone who can reach it (#7830)
• Long-Context Fast Tier Pricing - Added ultrafast and priority above-272k pricing columns so fast-tier requests over 272k tokens, including cache writes, bill at the published long-context rates (#7834, #7838)

🐞 Fixed

• Route parsing bug fix - Auth whitelist and temp-token scope checks fixes fasthttp routing bug
• Code Mode Auto-Execute Allow List - tools_to_execute and tools_to_auto_execute are now enforced at invocation time inside code mode, so indirect calls like getattr(server, name)(...) or a plugin tool rename cannot bypass them. Approved runs through /v1/mcp/tool/execute are bound only by tools_to_execute (#7833)
• OpenAI service_tier Fast Billing - Requests with service_tier fast now bill at the priority rates, and the tier is echoed back to clients (#7837)
• Gemini to OpenAI Fallback - Fallbacks from Gemini to OpenAI Responses now strip fields OpenAI rejects: item status, generated reasoning and function output IDs, function output name, and content signatures (#7835)
• Guardrail Redaction Alignment - Anthropic raw transform targets now match normalized guardrail ordinals when billing headers or MCP blocks are present, so redaction hits the right fields (#7808) …