Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 900+ contributors
TypeScript a2aai-agentsai-gatewayanthropicclaudeclaude-codeclinecodexcopilotcursordeepseekfree-aigeminikimillm-gatewaymcpopenaiopenai-proxyqwentoken-saver
export interface FeatureFlagDefinition {
key: string;
label: string;
description: string;
descriptionI18nKey: string;
category: "security" | "network" | "policies" | "runtime" | "cli" | "health";
defaultValue: string;
type: "boolean" | "enum";
enumValues?: string[];
requiresRestart: boolean;
warningLevel?: "info" | "caution" | "danger";
}
export const FEATURE_FLAG_DEFINITIONS: FeatureFlagDefinition[] = [
// ──────────────── Security (10) ────────────────
{
key: "REQUIRE_API_KEY",
label: "Require API Key",
description: "Require an API key for all incoming requests",
descriptionI18nKey: "featureFlagRequireApiKeyDescription",
category: "security",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "INPUT_SANITIZER_ENABLED",
label: "Input Sanitizer",
description: "Enable input sanitization for all requests",
descriptionI18nKey: "featureFlagInputSanitizerEnabledDescription",
category: "security",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "INJECTION_GUARD_MODE",
label: "Injection Guard Mode",
description: "Set the prompt injection guard mode",
descriptionI18nKey: "featureFlagInjectionGuardModeDescription",
category: "security",
defaultValue: "off",
type: "enum",
enumValues: ["off", "warn", "block", "redact"],
requiresRestart: false,
warningLevel: "info",
},
{
key: "PII_REDACTION_ENABLED",
label: "PII Redaction",
description: "Redact personally identifiable information from requests",
descriptionI18nKey: "featureFlagPiiRedactionEnabledDescription",
category: "security",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "PII_RESPONSE_SANITIZATION",
label: "PII Response Sanitization",
description: "Sanitize PII from provider responses",
descriptionI18nKey: "featureFlagPiiResponseSanitizationDescription",
category: "security",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "PII_RESPONSE_SANITIZATION_MODE",
label: "PII Response Sanitization Mode",
description:
"Mode for PII response sanitization: redact (replace PII), warn (log only), block (reject), off (disable)",
descriptionI18nKey: "featureFlagPiiResponseSanitizationModeDescription",
category: "security",
defaultValue: "redact",
type: "enum",
enumValues: ["redact", "warn", "block", "off"],
requiresRestart: false,
warningLevel: "info",
},
{
key: "OUTBOUND_SSRF_GUARD_ENABLED",
label: "SSRF Guard",
description:
"Legacy alias: a value saved on this flag's dashboard toggle is read before the environment; false, 0, no, or off in either turns the outbound URL guard's host checks off like OMNIROUTE_ALLOW_PRIVATE_PROVIDER_URLS",
descriptionI18nKey: "featureFlagOutboundSsrfGuardEnabledDescription",
category: "security",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "ALLOW_API_KEY_REVEAL",
label: "API Key Reveal",
description:
"Allow authenticated dashboard users to reveal stored API keys instead of only seeing masked values.",
descriptionI18nKey: "featureFlagAllowApiKeyRevealDescription",
category: "security",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "AUTH_LOG_INCLUDE_ACCOUNT_ID",
label: "Log Account IDs",
description:
'Include account prefix in AUTH log lines (e.g. "Using <provider> account: abc12345..."). ' +
"Disabled by default so account identifiers are redacted from shared/multi-tenant process logs. " +
"Independent from Debug Mode; flipping Debug Mode does not reveal this.",
descriptionI18nKey: "featureFlagAuthLogIncludeAccountIdDescription",
category: "security",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_OIDC_DISABLE_PASSWORD_LOGIN",
label: "Disable Password Login With OIDC",
description:
"When OIDC is enabled, disable password login so users can only authenticate via OIDC Single Sign-On. When disabled (default), both password login and OIDC are available.",
descriptionI18nKey: "featureFlagOidcDisablePasswordLoginDescription",
category: "security",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
// ──────────────── Network (7) ────────────────
{
key: "ENABLE_TLS_FINGERPRINT",
label: "TLS Fingerprint",
description: "Enable TLS fingerprint stealth mode",
descriptionI18nKey: "featureFlagEnableTlsFingerprintDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: true,
warningLevel: "info",
},
{
key: "AUDIO_REMOTE_PROVIDER_NODES",
label: "Remote Audio Provider Nodes",
description:
"Allow the /v1/audio/* routes to use OpenAI-compatible provider nodes hosted outside localhost. Off by default — routing audio to a remote host changes egress identity and must be an explicit operator decision. Loopback nodes are always allowed and unaffected.",
descriptionI18nKey: "settings.featureFlags.audioRemoteProviderNodes",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "RERANK_REMOTE_PROVIDER_NODES",
label: "Remote Rerank Provider Nodes",
description:
"Allow POST /v1/rerank (and the memory engine's rerank step, which calls it over loopback) to use OpenAI-compatible provider nodes hosted outside localhost — a LAN box or Tailscale peer running TEI, Infinity, vLLM, etc. Off by default — routing to a remote host changes egress identity and must be an explicit operator decision. Loopback nodes are always allowed and unaffected. Remote nodes must also pass the provider outbound URL policy.",
descriptionI18nKey: "settings.featureFlags.rerankRemoteProviderNodes",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "PROXY_AUTO_SELECT_ENABLED",
label: "Proxy Auto-Selection Fallback",
description:
"When no proxy is assigned to a connection, auto-select the first working proxy from the registry. Off by default — otherwise any single registry proxy becomes a global fallback for all traffic (#3332).",
descriptionI18nKey: "settings.featureFlags.proxyAutoSelectEnabled",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OMNIROUTE_CONTROL_PLANE_PROXY_DIRECT_FALLBACK",
label: "Control-Plane Proxy Direct Fallback",
description:
"Allow OAuth and provider validation flows to bypass a pinned proxy and connect directly when proxy reachability pre-checks fail. Off by default because this can change account egress IP.",
descriptionI18nKey: "featureFlagOmnirouteControlPlaneProxyDirectFallbackDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "NETWORK_ROTATION_SHARED_EGRESS_GUARD",
label: "Network Rotation Shared-Egress Guard",
description:
"On a network exception (timeout, connection refused/reset) for a multi-account rotation executor, when the failing account has no dedicated proxy, apply a short cooldown and skip other proxy-less accounts for the rest of the request instead of retrying each one. On by default (safe: no egress IP change, only reduces latency/cooldown risk on shared-egress accounts). Disable to restore immediate propagation on the first proxy-less throw.",
descriptionI18nKey: "featureFlagNetworkRotationSharedEgressGuardDescription",
category: "network",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "PROXY_SKIP_RECENTLY_FAILED",
label: "Skip Recently Failed Proxies",
description:
"Proxy pools and the per-account rotation of opencode stop re-serving a proxy that just failed (refused TCP probe, or a 429 received through it) for a per-process period that doubles on each repeat, up to a cap. No proxy status is written; with every candidate set aside the choice is unchanged. On by default: selection order is exactly the plain rotation only with PROXY_SKIP_RECENTLY_FAILED=false.",
descriptionI18nKey: "featureFlagProxySkipRecentlyFailedDescription",
category: "network",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "PROXY_POOL_SHARED_EGRESS_ORDER",
label: "Shared Egress Pool Order",
description:
"For providers whose quota is bucketed by egress address, rank a pool member sharing a recently refused member's observed egress address just below healthy members. Order only, never excluded. Needs PROXY_SKIP_RECENTLY_FAILED, which produces the refusal signal it reads. Off by default: selection order is exactly the plain rotation.",
descriptionI18nKey: "featureFlagProxyPoolSharedEgressOrderDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "ROTATION_ATTRIBUTION",
label: "Rotation Attribution Logging",
description:
"Opencode rotation records which account served or was skipped (masked ids only, never full account ids) and links proxy log entries to their request, so the operator can tell skipped accounts apart from unused ones. Off by default: no extra log lines, no extra columns written.",
descriptionI18nKey: "featureFlagRotationAttributionDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "PROXY_POOL_EGRESS_OBSERVATION",
label: "Proxy Pool Egress Observation",
description:
"Show, under a proxy pool in the dashboard, how many observed egress IPs served its members over the last 24 h, how many connections used them and the most seen behind one IP. Read-only, computed from the proxy log, never used for routing. Off by default: the pool editor is unchanged and the observation route answers null.",
descriptionI18nKey: "featureFlagProxyPoolEgressObservationDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "PROXY_OPERATOR_EGRESS_ENABLED",
label: "Proxy Operator Egress",
description:
"Accept operator-pushed dated observed addresses per pool member and merge them with the journal read for display and pool order. Off by default: the push route answers 404 and pool reads behave exactly as before.",
descriptionI18nKey: "featureFlagProxyOperatorEgressDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_RESPONSES_STALL_ROTATION",
label: "OpenCode Responses Stall Rotation",
description:
"For the OpenCode executor, watch the first body byte of a streamed Responses reply (window: RESPONSES_FIRST_BYTE_TIMEOUT_MS, default 15000). A 2xx Responses stream that stays silent past the window is treated as stalled: the account is cooled down and the request rotates to the next account once; a second stall fails fast. Off by default: stalled streams keep today's wait until the stream readiness timeout.",
descriptionI18nKey: "featureFlagOpencodeResponsesStallRotationDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_USER_BLOCKED_ROTATION",
label: "OpenCode user_blocked Rotation",
description:
"For the OpenCode executor, when an upstream answers 403 or 451 carrying a user_blocked refusal (not a geo block, not a Cloudflare fingerprint rejection), cool the refused account down and rotate to the next account at most once per request; a second refusal is returned as-is without a success mark. Off by default: routing around an upstream user block can look like evasion and spread the flag across the account fleet, so the refusal is returned unchanged unless the operator opts in.",
descriptionI18nKey: "featureFlagOpencodeUserBlockedRotationDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_TRANSIENT_FAILOVER_BACKOFF",
label: "OpenCode Transient Failover Backoff",
description:
"For the OpenCode multi-account rotation, pause before dispatching to the next account once two consecutive attempts failed with a transient upstream error (5xx or an empty 400 rejection). The pause starts at 1.5s, doubles per further consecutive failure, is capped at 6s per pause and 10s per request, is skipped when the client disconnects, and the failed response body is released before waiting. Off by default: failover stays immediate.",
descriptionI18nKey: "featureFlagOpencodeTransientFailoverBackoffDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "FLUSH_EMPTY_RETRY_ENABLED",
label: "Flush Empty Turn Retry",
description:
"On translated streaming turns, when the upstream turn carries no usable content (reasoning-only completion or zero valuable chunks), issue bounded retries through the normal credential path (up to `STREAM_RECOVERY.EMPTY_TURN_RETRY_MAX`) before anything is exposed to the client. Off by default: empty turns keep the current behavior (empty 200 or empty-content 502).",
descriptionI18nKey: "featureFlagFlushEmptyRetryEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "EMPTY_STREAM_SIBLING_RECOVERY_ENABLED",
label: "Empty Stream Sibling Recovery",
description:
"Claude/Codex streaming: after the first lifecycle event, briefly probe (up to 1s / 64 KiB) for usable output; a stream that ends with no content becomes an empty_content failure that spends the existing bounded retry on another eligible connection (no cooldown, no model lockout), and round-robin combos stop retrying once the client disconnects. Also treats refusal, pause_turn and model_context_window_exceeded as legitimate empty terminals inside the probe. Off by default: readiness, empty-content handling and combo retries stay exactly as before.",
descriptionI18nKey: "featureFlagEmptyStreamSiblingRecoveryEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "STREAM_LATE_OUTPUT_PROBE_ENABLED",
label: "Stream Late Output Probe",
description:
"Streaming output probe: after the first 64 forwarded chunks keep sampling one chunk in 16 for the first visible output, so output that starts late still marks the stream as having emitted output. A lone 5xx after that point then stays request-scoped instead of recording a model-only lockout, and firstOutputMs is back-filled to the first output chunk of the current sampling gap. Off by default: only the first 64 forwarded chunks are probed, exactly as before.",
descriptionI18nKey: "featureFlagStreamLateOutputProbeEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_RATE_LIMITED_429_EARLY_STOP",
label: "OpenCode Rate-Limited 429 Early Stop",
description:
"For the OpenCode multi-account rotation, stop the account wave at the first 429 classified as a real rate limit (a parseable Retry-After header, or a body naming a rate/usage limit) and return that upstream 429 unchanged (status, body, Retry-After and quota headers), instead of trying every remaining account. Unclassified 429s keep rotating. Off by default: the free tier is limited per egress IP (#9611), so every 429 rotates to the next account, and an exhausted wave returns the last upstream 429.",
descriptionI18nKey: "featureFlagOpencodeRateLimited429EarlyStopDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_PARK_AND_RESUME",
label: "OpenCode 429 Park And Resume",
description:
"For the OpenCode multi-account rotation, park the request after repeated transient 429s (or a fresh pool-strain marker) with a heartbeat, then replay one capped leg of up to 3 sequential accounts instead of fanning out the whole fleet. Off by default: every 429 rotates to the next account exactly as before.",
descriptionI18nKey: "featureFlagOpencodeParkAndResumeDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "STREAM_READINESS_STALL_RETRY",
label: "Stream Readiness Stall Retry",
description:
"For streaming chat requests, when the first upstream body stalls before producing a usable event, issue one bounded second attempt through the same routing path with the same readiness budget and no account penalty. Off by default: a stalled first body fails the request without a retry.",
descriptionI18nKey: "featureFlagStreamReadinessStallRetryDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_POOL_RESELECT",
label: "OpenCode 429 Pool Reselect",
description:
"For the OpenCode multi-account rotation, after a 429 from an egress-bucketed provider on a proxy-less account under an ambient pool context, ask the connection pool for another member for the next attempt instead of retrying the same egress address. Orders, never excludes: an exhausted pool keeps the current behavior. Off by default: every 429 rotates to the next account exactly as before.",
descriptionI18nKey: "featureFlagOpencodePoolReselectDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_V2_CLIENT_FINGERPRINT_ENABLED",
label: "OpenCode v2 Client Fingerprint",
description:
'Send the OpenCode v2 CLI client fingerprint on OpenCode Zen/Go requests (#16571): User-Agent opencode/latest/<2.x>/cli and x-opencode-client cli, a client User-Agent kept only when it is a 2.x OpenCode client, and on gated free-tier requests no x-opencode-request, the session mirrored onto x-opencode-session-id/x-session-affinity/x-session-id, the v2 CLI read/shell schemas on injected placeholder tools and a tool_choice of "none" dropped. Off by default: every OpenCode request keeps the current identity, headers and body byte for byte.',
descriptionI18nKey: "featureFlagOpencodeV2ClientFingerprintEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_ZEN_EGRESS_IP_LOCK_ENABLED",
label: "OpenCode Zen Egress-IP Lockout",
description:
"Add OpenCode Zen (opencode-zen) to the default egress-IP lockout family, so a 429 on one Zen connection also cools every sibling Zen connection that shares its egress IP. Off by default: Zen is keyed and each connection cools down on its own, exactly as before. An explicit OMNIROUTE_EGRESS_IP_LOCK_PROVIDERS value always wins over this flag.",
descriptionI18nKey: "featureFlagOpencodeZenEgressIpLockEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_FULL_TEXT_ERROR_RULES_ENABLED",
label: "OpenCode Full-Text Error Rules",
description:
'Let the built-in OpenCode error rules (opencode-zen, opencode-go, opencode-cli) match on the full upstream error text, so body markers such as organization_quota_exceeded or "monthly usage limit reached" classify a 429 as quota exhausted with a connection-scoped lock. Off by default: those rules only see the structured error code and such a 429 stays a short rate limit, exactly as before.',
descriptionI18nKey: "featureFlagOpencodeFullTextErrorRulesEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CLOUDFLARE_AI_FULL_TEXT_ERROR_RULES_ENABLED",
label: "Cloudflare Workers AI Full-Text Error Rules",
description:
'Let the built-in Cloudflare Workers AI error rule match on the full upstream error text, so the 429 "daily free allocation" body is classified as quota exhausted instead of a short rate limit. Off by default: the rule only sees the structured error code, exactly as before.',
descriptionI18nKey: "featureFlagCloudflareAiFullTextErrorRulesEnabledDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "MITM_DISABLE_TLS_VERIFY",
label: "Disable TLS Verify (MITM)",
description: "Disable TLS certificate verification for MITM proxy",
descriptionI18nKey: "featureFlagMitmDisableTlsVerifyDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: true,
warningLevel: "danger",
},
{
key: "OMNIROUTE_ALLOW_PRIVATE_PROVIDER_URLS",
label: "Allow Private Provider URLs",
description:
"Turns off the outbound URL guard's host checks, cloud-metadata block included, on provider URL validation, model discovery, provider-node base URLs, and the proxy-fallback test, and allows private webhook targets. Local and LAN URLs already pass on the validation, discovery, and provider-node paths by default; the proxy-fallback test and webhook targets check only this flag",
descriptionI18nKey: "featureFlagOmnirouteAllowPrivateProviderUrlsDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OMNIROUTE_ALLOW_LOCAL_PROVIDER_URLS",
label: "Allow Local Provider URLs",
description:
"Allow provider URLs on local/private addresses (127.0.0.1, localhost, LAN). On by default (OmniRoute is local-first): the guard then blocks cloud-metadata endpoints (all of 169.254.0.0/16 plus the known metadata hostnames). Disable for strict public-only blocking: private and loopback hosts are blocked too",
descriptionI18nKey: "featureFlagOmnirouteAllowLocalProviderUrlsDescription",
category: "network",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "ENABLE_CC_COMPATIBLE_PROVIDER",
label: "CC Compatible Provider",
description: "Enable Claude Code compatible provider mode",
descriptionI18nKey: "featureFlagEnableCcCompatibleProviderDescription",
category: "network",
defaultValue: "false",
type: "boolean",
requiresRestart: true,
warningLevel: "info",
},
// ──────────────── Policies (7) ────────────────
{
key: "TOOL_POLICY_MODE",
label: "Tool Policy Mode",
description: "Set the tool use policy enforcement mode",
descriptionI18nKey: "featureFlagToolPolicyModeDescription",
category: "policies",
defaultValue: "disabled",
type: "enum",
enumValues: ["disabled", "warn", "block"],
requiresRestart: false,
warningLevel: "info",
},
{
key: "RATE_LIMIT_AUTO_ENABLE",
label: "Rate Limit Auto-Enable",
description: "Automatically enable rate limiting based on usage patterns",
descriptionI18nKey: "featureFlagRateLimitAutoEnableDescription",
category: "policies",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "DISABLE_CONTEXT_WINDOW_CHECKS",
label: "Disable Context Window Checks",
description:
"Skip OmniRoute's local context-window and max-input-token check for direct single-model requests. Upstream providers remain responsible for enforcing their actual limits. Off by default.",
descriptionI18nKey: "featureFlagDisableContextWindowChecksDescription",
category: "policies",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "CAPABILITY_FILTER_ENABLED",
label: "Capability Filter",
description:
"Reject requests before dispatch when the target model lacks required capabilities (vision, tools, structured output, context window). Protects direct single-provider requests that bypass the combo-layer compatibility filter.",
descriptionI18nKey: "featureFlagCapabilityFilterEnabledDescription",
category: "policies",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "USAGE_LIMIT_IGNORE_UNPRICED",
label: "Ignore Unpriced Usage in USD Quotas",
description:
"Count usage of models that have no price as $0 in per-key USD usage quotas instead of treating the quota as exceeded. Off by default: an unpriced model or routing alias can hide real spend, so the quota fails closed.",
descriptionI18nKey: "featureFlagUsageLimitIgnoreUnpricedDescription",
category: "policies",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "RADAR_ENABLED",
label: "Radar",
description:
"Enable the OmniRoute Radar module (catalog feed screens and sync). Off by default; enabling only unlocks the UI — data sync remains a separate opt-in.",
descriptionI18nKey: "featureFlagRadarEnabledDescription",
category: "policies",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "MODEL_ENDPOINT_CAPABILITY_GATE_ENABLED",
label: "Model Endpoint Capability Gate",
description:
"Enforce trusted model endpoint tags (declared by the provider catalog, inferred from a specialty model id, or set by the operator): reject a request sent to an endpoint the model does not serve with HTTP 400 before dispatch on /v1/chat/completions, /v1/embeddings, /v1/rerank, /v1/images/generations, /v1/videos/generations, /v1/music/generations and /v1/audio/*; hide models whose tagged endpoints have no OmniRoute adapter from /v1/models; and answer the dashboard model test with 422 for such models. Off by default: endpoint tags stay informational and requests are dispatched exactly as before.",
descriptionI18nKey: "featureFlagModelEndpointCapabilityGateEnabledDescription",
category: "policies",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
// ──────────────── Runtime (18) ────────────────
{
key: "UNIVERSAL_CONTEXT_HANDOFF_ENABLED",
label: "Universal Context Handoff",
description:
"Generate and inject conversation summaries when combo routing switches models. Disable to treat model switches independently and prevent background handoff requests for all existing and future combos.",
descriptionI18nKey: "featureFlagUniversalContextHandoffEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "REASONING_REPLAY_ENABLED",
label: "Reasoning Replay",
description:
"Cache and replay model reasoning across multi-turn conversations. Disable to stop storing and re-injecting reasoning.",
descriptionI18nKey: "featureFlagReasoningReplayEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "RESPONSES_PASSTHROUGH_DROP_COMMENTARY",
label: "Drop Responses Commentary",
description:
"Drop internal commentary-phase output items from Responses API passthrough streams before forwarding to clients. Disable to receive raw upstream commentary.",
descriptionI18nKey: "featureFlagResponsesPassthroughDropCommentaryDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_MCP_ENFORCE_SCOPES",
label: "MCP Enforce Scopes",
description: "Enforce scope restrictions on MCP tool access",
descriptionI18nKey: "featureFlagOmnirouteMcpEnforceScopesDescription",
category: "runtime",
// Ships off: the gate rejects a caller that sends no scopes at all, so turning it on
// is an operator decision (.env.example has shipped `=false` since the gate landed).
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OMNIROUTE_MCP_COMPRESS_DESCRIPTIONS",
label: "MCP Compress Descriptions",
description: "Compress MCP tool descriptions to reduce token usage",
descriptionI18nKey: "featureFlagOmnirouteMcpCompressDescriptionsDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_ENABLE_RUNTIME_BACKGROUND_TASKS",
label: "Runtime Background Tasks",
description: "Enable background task processing at runtime",
descriptionI18nKey: "featureFlagOmnirouteEnableRuntimeBackgroundTasksDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_DISABLE_BACKGROUND_SERVICES",
label: "Disable Background Services",
description: "Disable all background services (quota refresh, sync, etc)",
descriptionI18nKey: "featureFlagOmnirouteDisableBackgroundServicesDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: true,
warningLevel: "caution",
},
{
key: "OMNIROUTE_RTK_TRUST_PROJECT_FILTERS",
label: "RTK Trust Project Filters",
description: "Trust project-level filters from RTK without validation",
descriptionI18nKey: "featureFlagOmnirouteRtkTrustProjectFiltersDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OMNIROUTE_ENABLE_LIVE_WS",
label: "Live Dashboard WebSocket",
description:
"Start the real-time dashboard WebSocket server on import (port 20132, loopback-bound by default). Default: enabled. Set to '0' or 'false' to disable. LAN exposure requires LIVE_WS_HOST=0.0.0.0 + LIVE_WS_ALLOWED_ORIGINS.",
descriptionI18nKey: "featureFlagOmnirouteEnableLiveWsDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: true,
warningLevel: "info",
},
{
key: "OMNIROUTE_CODEX_WS_ENABLED",
label: "Codex Responses WebSocket",
description:
"Allow Codex to use the Responses-over-WebSocket transport (the codex CLI WS endpoint and codexTransport=websocket). When off, Codex falls back to HTTP Responses.",
descriptionI18nKey: "featureFlagOmnirouteCodexWsEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_CODEX_APP_SERVER_ENABLED",
label: "Codex App-Server Transport",
description:
"Allow Codex to use the local app-server WebSocket JSON-RPC transport (codexTransport=app-server). When off, connections opted into app-server fall back to Codex's other transports.",
descriptionI18nKey: "featureFlagOmnirouteCodexAppServerEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_EMERGENCY_FALLBACK",
label: "Emergency Fallback",
description: "Route budget-exhausted requests to the emergency free fallback provider/model.",
descriptionI18nKey: "featureFlagOmnirouteEmergencyFallbackDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "STREAM_RECOVERY_ENABLED",
label: "Stream Recovery",
description:
"Enable transparent early retry for truncated upstream SSE streams before any response bytes reach the client.",
descriptionI18nKey: "featureFlagStreamRecoveryEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "STREAM_RECOVERY_MIDSTREAM_ENABLED",
label: "Mid-Stream Continuation",
description:
"Allow stream recovery to re-request and stitch a response after bytes have already reached the client.",
descriptionI18nKey: "featureFlagStreamRecoveryMidstreamEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "STREAM_RECOVERY_TOOLCALL_ORDER_FIX",
label: "Tool-Call-Safe Continuation",
description:
"Make mid-stream continuation tool-call safe: never resume a cut stream once a tool call was emitted (in flight or already finished with finish_reason tool_calls), and close after one empty continuation instead of spending the whole budget. Off: release behavior.",
descriptionI18nKey: "featureFlagStreamRecoveryToolcallOrderFixDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "STREAM_EARLY_EOF_SIBLING_FAILOVER_ENABLED",
label: "Early-EOF Sibling Failover",
description:
"Fail over once to a sibling connection when an SSE stream closes before emitting any useful frame and the bounded same-connection retry is spent; with no usable sibling the original STREAM_EARLY_EOF 502 is returned. Off by default: early-EOF stays terminal after the same-connection retry.",
descriptionI18nKey: "featureFlagStreamEarlyEofSiblingFailoverEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "MODEL_CATALOG_INCLUDE_NAMES",
label: "Model Catalog Names",
description:
"Include display-friendly name fields in /v1/models responses. Disable for clients that expect model IDs only.",
descriptionI18nKey: "settings.featureFlags.modelCatalogIncludeNames",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "MODELS_CATALOG_PREFIX_MODE",
label: "Models Catalog Prefix Mode",
description:
"Controls how model IDs are prefixed in /v1/models. 'dual' (default) emits both alias and canonical provider-id prefixes for backward compatibility. 'alias' emits only the short alias prefix (e.g. ds-web/model, not deepseek-web/model). 'canonical' emits only the full provider-id prefix.",
descriptionI18nKey: "featureFlagModelsCatalogPrefixModeDescription",
category: "runtime",
defaultValue: "dual",
type: "enum",
enumValues: ["dual", "alias", "canonical"],
requiresRestart: false,
warningLevel: "info",
},
{
key: "ARENA_ELO_SYNC_ENABLED",
label: "Arena ELO Sync",
description: "Enable periodic Arena AI leaderboard ELO sync for model intelligence rankings.",
descriptionI18nKey: "featureFlagArenaEloSyncEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "EXPOSE_CC_DISCOVERY_ALIASES",
label: "Claude Code Discovery Aliases",
description:
"Advertise claude/<provider>/<model> mirror ids on /v1/models so Claude Code gateway model discovery lists non-Claude models. Warning: doubles catalog entries for all clients when enabled globally.",
descriptionI18nKey: "featureFlagExposeCcDiscoveryAliasesDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "NO_THINKING_ALIAS_ENABLED",
label: "No-Thinking Model Aliases",
description:
"Master switch for the no-think/<provider>/<model> gateway aliases. On (default): /v1/models advertises a no-thinking variant for every eligible thinking-capable Claude model, and a no-think/ id sent on a request resolves back to the real model with reasoning suppressed. Off: no variants are advertised and a no-think/ id is treated like any other unknown model id. The per-model ModelSpec.noThinkingAlias opt-in/opt-out still applies while this is on.",
descriptionI18nKey: "featureFlagNoThinkingAliasEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_DISABLE_THINKING_LEVEL_VARIANTS",
label: "Disable Thinking Level Variants",
description:
"Disable the generation of thinking level variants (e.g. -low, -medium, -high) in the /v1/models catalog.",
descriptionI18nKey: "featureFlagOmnirouteDisableThinkingLevelVariantsDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_CHAT_VIRTUAL_LANES",
label: "Adaptive Virtual Admission Lanes",
description:
"Enable per-tenant adaptive virtual admission lanes for provider dispatch (#9654): one tenant's burst no longer 503s another. The OMNIROUTE_CHAT_VIRTUAL_LANES env var wins over this dashboard override; changes take effect at server restart.",
descriptionI18nKey: "featureFlagChatVirtualLanesEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: true,
warningLevel: "info",
},
{
key: "EXPOSE_FUNCTIONAL_GATEWAY_MIRRORS",
label: "Functional Gateway Mirrors",
description:
"Advertise <gateway-alias>/<model> mirror ids on /v1/models for models whose canonical owner has no active credential but a passthrough gateway with an active credential routes them. Warning: adds catalog entries for all clients when enabled globally.",
descriptionI18nKey: "featureFlagExposeFunctionalGatewayMirrorsDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "NEWAPI_AGGREGATOR_BALANCE",
label: "New-API Aggregator Balance",
description:
"Enable balance detection for New-API / One-API / Sub2API aggregator compatible nodes. When enabled, compatible nodes with the aggregator flag set will report their balance in the dashboard and quota-preflight routing.",
descriptionI18nKey: "featureFlagNewApiAggregatorBalanceDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "SERVER_OWNED_TOOL_LOOP_ENABLED",
label: "Server-Owned Tool Loop",
description:
"Continue non-streaming server-owned tool calls until the model returns a client-usable response.",
descriptionI18nKey: "featureFlagServerOwnedToolLoopDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "SEARCH_STATS_HIDE_DELETED_CONNECTIONS",
label: "Hide Deleted Search Connections",
description:
"Search stats and recent searches only count providers that still have a live connection (keyless providers such as duckduckgo-free always count). Off keeps every retained search row with a provider id.",
descriptionI18nKey: "featureFlagSearchStatsHideDeletedConnectionsDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "FREE_BADGE_REQUIRES_PROVIDER_FREE_TIER",
label: "Strict Free Badge",
description:
"Dashboard provider pages: show the Free badge only on signals the provider honors — drops the display-name heuristic, non-boolean free fields and :free suffixes on registered providers without a documented free tier. Off keeps the historical badge rule.",
descriptionI18nKey: "featureFlagFreeBadgeRequiresProviderFreeTierDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "RETRY_AFTER_PROVENANCE_ENABLED",
label: "Retry-After Provenance",
description:
"On aggregated 429/503 unavailable responses, omit Retry-After when no concrete future retry time is known (instead of a synthetic 1s), add error.retry_after_provenance (signal | none), and let combo drain paths read prose retry hints from JSON and plain-text upstream bodies.",
descriptionI18nKey: "featureFlagRetryAfterProvenanceEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "PROTECTED_PRIORITY_INFRA_502_ENABLED",
label: "Protected-Priority Infra Stops as 502",
description:
"When a priority combo target marked fallback-only-on-quota-exhaustion stops the combo for a cause that is provably not quota (provider circuit breaker open, predictive latency skip), answer 502 instead of the quota-looking 503. Lockout, cooldown, unavailable, exhaustion and concurrency-cap stops keep 503.",
descriptionI18nKey: "featureFlagProtectedPriorityInfra502EnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "COMBO_EMPTY_POOL_NO_ELIGIBLE_TARGETS_ENABLED",
label: "Auto Combo: Empty Pool Answers 503",
description:
'When an auto combo\'s candidate pool is already empty before the tool-calling pre-filter runs (every target skipped by cooldowns, lockouts or exhaustion), answer 503 no_eligible_targets instead of the legacy 400 "No target supports tool calling" capability_mismatch. Off by default: the empty pool keeps the legacy 400.',
descriptionI18nKey: "featureFlagComboEmptyPoolNoEligibleTargetsEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "MISTRAL_AMBIGUOUS_401_SOFT_LOCKOUT",
label: "Mistral Ambiguous 401 Soft Lockout",
description:
'A bare Mistral 401 ({"detail":"Unauthorized"}, no explicit auth signal) is byte-identical for a revoked key and for exhausted quota. When enabled, such a 401 cools the connection down instead of parking it as expired, up to 3 times within an hour; the next one still parks it as expired, so a revoked key converges. Off by default: every bare Mistral 401 parks the connection as expired, as before.',
descriptionI18nKey: "featureFlagMistralAmbiguous401SoftLockoutDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "BATCH_AND_FILE_AUTO_CLEANUP_ENABLED",
label: "Batch & File Auto-Cleanup",
description:
"Let the automatic cleanup sweep delete terminal (completed/failed/cancelled/expired) Batch API jobs older than OMNIROUTE_BATCH_RETENTION_DAYS, along with their per-line checkpoints, and clear the BLOB content of uploaded files past their own expires_at. Off by default: every existing install keeps this data exactly as before until an operator opts in. The operator-triggered DELETE /api/v1/batches/delete-completed route is unaffected either way -- it is a separate, unconditional public API contract.",
descriptionI18nKey: "featureFlagBatchAndFileAutoCleanupEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "danger",
},
{
key: "ANTIGRAVITY_ACCOUNT_LEASE_ENABLED",
label: "Antigravity Account Lease",
description:
"Reserve the selected Antigravity account for the streaming lifecycle of the request that picked it, so a concurrent retry or the credential handoff cannot re-pick an account already committed to an in-flight stream. The reservation is scoped to (connection, callable upstream model), so one account can still serve two different models at once. When every eligible account is already leased for that model, the request returns a structured 503 POOL_BUSY with a bounded Retry-After instead of piling onto a busy account. Off by default: account selection stays exactly as before, and no reservation is taken.",
descriptionI18nKey: "featureFlagAntigravityAccountLeaseEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "COMBO_AUTO_PRUNE_STALE_STEPS",
label: "Auto-Prune Stale Combo Steps",
description:
"After a successful model sync against an authoritative live catalog, remove combo steps pinned to models the catalog no longer lists, with one audit entry per removed step. Never prunes on a failed, degraded or free-only sync, and never empties a combo. Off by default: stale steps are only flagged.",
descriptionI18nKey: "featureFlagComboAutoPruneStaleStepsDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CLAUDE_OAUTH_SUMMARIZED_THINKING_UNREDACTED",
label: "Claude OAuth Summarized Thinking",
description:
'Let a Claude OAuth request that asks for readable thinking (thinking.display "summarized") drop the redact-thinking-2026-02-12 Anthropic beta, which otherwise makes Anthropic return empty, signature-only thinking blocks. Off by default: that beta is part of the Claude Code OAuth wire fingerprint, so OAuth requests keep it unchanged. API-key Claude requests already honor summarized thinking without this flag; OMNIROUTE_REDACT_THINKING=false also reaches OAuth only when this flag is on.',
descriptionI18nKey: "featureFlagClaudeOauthSummarizedThinkingUnredactedDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "KIRO_TOOL_CALL_REPAIR_ENABLED",
label: "Kiro Tool Call Repair",
description:
"When a streaming Kiro response emits a malformed tool_call wrapper (missing input.name or input.arguments) before any visible output, suppress it and re-send the request once with a corrective instruction instead of ending the turn with invalid_kiro_tool_call. The retry costs one extra Kiro request; a second failure ends the turn with kiro_tool_call_repair_retry_failed. On by default. Set to false to fail the turn on the first malformed wrapper.",
descriptionI18nKey: "featureFlagKiroToolCallRepairEnabledDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "MODELS_DEV_EFFORT_TIERS_ENABLED",
label: "models.dev Effort Tiers",
description:
'Use the reasoning effort tiers models.dev publishes (narrowed to the known effort vocabulary) when neither an operator override nor the provider registry declares any, and treat a declared empty tier list as "no tiers" in /v1/models effort_tiers and combo tier intersections. Off by default: effort_tiers and resolved supportedThinkingEfforts stay exactly as before.',
descriptionI18nKey: "featureFlagModelsDevEffortTiersEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "LMARENA_SYNTHESIZED_CX_ENABLED",
label: "Arena Synthesized cx Telemetry",
description:
"Attach a synthesized programmatic cx client-interaction blob (the value arena.ai itself sends for a page with no recorded interaction) to every Arena (lmarena) create-evaluation request that has no operator-supplied providerSpecificData.cx. Off by default: the cx field is an Arena anti-automation signal, so OmniRoute only forwards a value the operator captured from a real browser session (providerSpecificData.cx) and otherwise sends no cx at all. Unproven upstream benefit; enable only if Arena starts rejecting requests without cx.",
descriptionI18nKey: "featureFlagLmarenaSynthesizedCxEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "NATIVE_CODEX_REACTIVE_COMPACTION_ENABLED",
label: "Native Codex Reactive Compaction",
description:
"Let the reactive (threshold) and last-resort context compaction passes also rewrite native Codex passthrough histories (POST /v1/responses to the codex provider) when they exceed the context budget, using the Responses input[] adapter. Off by default: native Codex histories are owned by the Codex client and are never rewritten by compaction; oversized requests keep failing the context budget check exactly as before. Requires global compression to be on and the model not to be excluded.",
descriptionI18nKey: "featureFlagNativeCodexReactiveCompactionEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "QUOTA_WINDOW_BOUNDED_PARK",
label: "Quota Window: Honour Bounded Park",
description:
"Let the per-window quota policy (session/weekly thresholds used by account selection) honour the bounded exhaustion park: once the capped park deadline has passed, or the cached quota observation is older than the park cap, the window stops excluding the account until a fresh quota fetch. Off by default: a cached window keeps blocking on its last reported percentage until its own reset time.",
descriptionI18nKey: "featureFlagQuotaWindowBoundedParkDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "IMAGE_GENERATION_LONG_JOBS",
label: "Image Generation: Long Jobs",
description:
"For OpenAI-compatible image generation: use the connection's timeout (providerSpecificData.timeoutMs), else the central upstream fetch timeout (FETCH_TIMEOUT_MS, then REQUEST_TIMEOUT_MS, default 600 s) instead of the fixed 120 s image budget; end a request the caller cancelled with 499; and let the Playground receive long image jobs as an SSE stream with keepalive heartbeats. Off by default: the fixed budget applies and caller cancellation is not propagated. The deadline covering the response body read is always on.",
descriptionI18nKey: "featureFlagImageGenerationLongJobsDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CATALOG_MIRROR_REFINEMENTS_ENABLED",
label: "Catalog Mirror Refinements",
description:
"Refine the claude/ discovery and functional-gateway mirrors on /v1/models. Off by default: both mirror passes behave exactly as before. On: claude/ discovery mirrors also cover listed reasoning-effort variants (-low/-medium/-high/-xhigh), skip built-in auto/* ids the request path cannot materialize and skip an alias already listed; functional-gateway mirrors skip combos and OmniRoute-synthesized ids (claude/, no-think/), take the owner from owned_by, never use the owner as its own gateway, skip gateways disabled in blockedProviders and models outside a gateway's authoritative live catalog.",
descriptionI18nKey: "featureFlagCatalogMirrorRefinementsEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
// ──────────────── CLI (5) ────────────────
{
key: "CLI_COMPAT_ALL",
label: "CLI Compat All",
description: "Enable compatibility mode for all CLI clients",
descriptionI18nKey: "featureFlagCliCompatAllDescription",
category: "cli",
defaultValue: "false",
type: "boolean",
requiresRestart: true,
warningLevel: "info",
},
{
key: "MODEL_ALIAS_COMPAT_ENABLED",
label: "Model Alias Compat",
description: "Enable model alias compatibility layer",
descriptionI18nKey: "featureFlagModelAliasCompatEnabledDescription",
category: "cli",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "PRICING_SYNC_ENABLED",
label: "Pricing Sync",
description:
"Enable automatic pricing data synchronization (requires the PRICING_SYNC_ENABLED environment variable to be set to true)",
descriptionI18nKey: "featureFlagPricingSyncEnabledDescription",
category: "cli",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_AUTO_SYNC_CODEX_PROFILES",
label: "Auto-Sync Codex Profiles",
description:
"After a provider model sync, automatically (re)write ~/.codex/*.config.toml profile files from the live catalog. Never changes the active/default Codex config. Off by default.",
descriptionI18nKey: "featureFlagOmnirouteAutoSyncCodexProfilesDescription",
category: "cli",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OMNIROUTE_AUTO_SYNC_CLAUDE_PROFILES",
label: "Auto-Sync Claude Code Profiles",
description:
"After a provider model sync, automatically (re)write ~/.claude/profiles/'<name>'/settings.json Claude Code profiles from the live catalog. Never changes the active/default Claude config. Off by default.",
descriptionI18nKey: "featureFlagOmnirouteAutoSyncClaudeProfilesDescription",
category: "cli",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
// ──────────────── Health (3) ────────────────
{
key: "OMNIROUTE_DISABLE_LOCAL_HEALTHCHECK",
label: "Disable Local Health Check",
description: "Disable the local instance health check endpoint",
descriptionI18nKey: "featureFlagOmnirouteDisableLocalHealthcheckDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "OMNIROUTE_DISABLE_TOKEN_HEALTHCHECK",
label: "Disable Token Health Check",
description: "Disable the token validation health check",
descriptionI18nKey: "featureFlagOmnirouteDisableTokenHealthcheckDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "SKILLS_SANDBOX_NETWORK_ENABLED",
label: "Skills Sandbox Network",
description: "Enable network access in the skills sandbox environment",
descriptionI18nKey: "featureFlagSkillsSandboxNetworkEnabledDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "PROXY_HEALTH_BLOCKED_RESETS_STREAK",
label: "Proxy Health: Refusal Resets Failure Streak",
description:
"In the proxy health sweep, let a probe the target refused (401/403/429: the proxy relayed, the destination refused this egress IP) reset the proxy's consecutive-failure streak, like a served probe. Off by default: a refusal stays neutral and keeps the streak (#10654). A 5xx stays inconclusive either way, and a refusal never removes, disables or re-activates a proxy.",
descriptionI18nKey: "featureFlagProxyHealthBlockedResetsStreakDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "GROK_SUBSCRIPTION_IMAGES_ENABLED",
label: "Grok Subscription Images",
description:
"Register xai-oauth (xao) and grok-cli image routes and map OpenAI quality high/hd to xAI medium. Off by default: the API-key xAI image path stays on the existing OpenAI-compatible request and the subscription routes are not registered.",
descriptionI18nKey: "featureFlagGrokSubscriptionImagesEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "XAI_OAUTH_LIVE_MODEL_DISCOVERY",
label: "xAI OAuth Live Model Discovery",
description:
"Fetch the live xAI model catalog for xai-oauth connections from https://api.x.ai/v1/models using the OAuth bearer token, instead of the frozen static seed. On by default. Set the flag to false to keep serving the static seed. HTTP failures fall back to the seed in the discovery route; the flag getter itself does not issue HTTP.",
descriptionI18nKey: "featureFlagXaiOauthLiveModelDiscoveryDescription",
category: "runtime",
defaultValue: "true",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "ANTHROPIC_COMPATIBLE_MID_CONVERSATION_SYSTEM",
label: "Anthropic-Compatible Mid-Conversation System Turns",
description:
"On plain anthropic-compatible-* relays (claude → claude passthrough), keep mid-conversation role:system turns inside messages[] instead of hoisting them into the top-level system parameter, so the cached prompt prefix stays stable across turns. Applies only when the client itself sent the mid-conversation-system-2026-04-07 anthropic-beta (forwarded upstream as-is). Off by default: every system turn is hoisted exactly as before, because relays that do not support the beta reject role:system messages with 400.",
descriptionI18nKey: "featureFlagAnthropicCompatibleMidConversationSystemDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "DB_HEALTHCHECK_STARTUP_DEFERRED_ENABLED",
label: "DB Health Check: Defer Startup Scan",
description:
"Run the startup DB integrity/health check after the server starts accepting requests (via setImmediate) instead of blocking startup until it completes. Off by default: startup blocks on the check exactly like before #13717, so a corrupt database is still caught before the first request is served. On: startup returns immediately and the check (now bounded/paged and, for a real file-backed DB, isolated in a cancellable child process) runs right after.",
descriptionI18nKey: "featureFlagDbHealthcheckStartupDeferredEnabledDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "OPENCODE_PUBLISHED_CATALOG_ROUTING_ENABLED",
label: "OpenCode Published Catalog Routing",
description:
"Let OpenCode Zen and OpenCode Go models without an explicit registry endpoint take their request surface (Responses for @ai-sdk/openai, Messages for @ai-sdk/anthropic) from the published models.opencode.ai catalog, read in the background every 6 h. Off by default: the endpoint comes from the registry heuristics exactly as before and the catalog is never fetched.",
descriptionI18nKey: "featureFlagOpencodePublishedCatalogRoutingEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "info",
},
{
key: "AUTO_REQUEST_RESOLVED_CAPABILITIES",
label: "Auto Requests: Resolved Capability Weights",
description:
"Let auto/* requests (resolveModelOrError and the auto-routing wrapper) build the candidate pool with resolved model capabilities (vision, reasoning, thinking) and observed quality scores, read in one bulk pass like the /v1/models catalog. Off by default: request-time snapshot weights stay exactly as before (capabilities resolve on demand and do not feed the weights). On: those signals feed the snapshot weights, so candidate weights, and therefore selection, can change.",
descriptionI18nKey: "featureFlagAutoRequestResolvedCapabilitiesDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "AUTO_COMBO_KEYLESS_TRANSPORT_GATE",
label: "Auto-Combo: Keyless Transport Gate",
description:
"Hold keyless (no-auth) providers whose transport needs a local browser (browserBacked) or a locally installed CLI (isLocalCli) out of the auto/* candidate pool, so a hung member cannot stall an auto request until the per-target timeout (#17036). Off by default: every no-auth provider with emulated tool calling stays in the pool exactly as before. Direct <alias>/<model> calls work either way.",
descriptionI18nKey: "featureFlagAutoComboKeylessTransportGateDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CODEBUDDY_CN_INLINE_REASONING",
label: "CodeBuddy CN: Inline Reasoning Replay",
description:
"For codebuddy-cn thinking models (the DeepSeek-V4 / Kimi-K2 families that need reasoning replay), inline each replayed assistant reasoning_content as a visible <thought> block at the front of the message content, so it survives a gateway that drops the reasoning_content field (#13632). Off by default: codebuddy-cn requests are sent unchanged; the reasoning_content field itself is always kept.",
descriptionI18nKey: "featureFlagCodebuddyCnInlineReasoningDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "PROVIDER_SAFETY_STOPS_TERMINAL",
label: "Provider Safety Stops: Terminal",
description:
"Keep an explicit provider safety stop (misalignment_policy_violation, content_policy_violation, or the exact legacy 'blocked by our safety systems' rejection labelled provider_safety_rejection) as a terminal, request-scoped refusal: its code is preserved through HTTP, Responses, WebSocket and completed-stream errors, and a combo stops at it instead of retrying other targets, cancelling admitted siblings (#17049). Off by default: these errors keep their existing handling and combo routing is unchanged.",
descriptionI18nKey: "featureFlagProviderSafetyStopsTerminalDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CREDENTIAL_HEALTH_RETEST_AUTO_DISABLED",
label: "Credential Health: Retest Auto-Disabled Connections",
description:
"Let the credential health sweep retest connections the gateway switched off by itself (marked autoDisabledAt; at most once per 6 h each and 5 per sweep) and switch a connection back on when its retest passes (#16190). Off by default: switched-off connections stay off until an operator re-enables or re-tests them, exactly like before. Operator-disabled connections and connections switched off after a permanent ban are never retested.",
descriptionI18nKey: "featureFlagCredentialHealthRetestAutoDisabledDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CREDENTIAL_GATE_AUTH_DIAGNOSES_ONLY",
label: "Credential Gate: Auth Diagnoses Only",
description:
"Narrow the combo credential gate so a cached failed credential probe blocks a connection only when its diagnosis proves bad credentials (upstream_auth_error, auth_missing, token_expired, token_refresh_failed, reauth_required, account_deactivated, auth_required). Inconclusive failures (timeouts, rate limits, outages, unsupported probes and untyped scheduler errors) are then treated as unknown and the connection keeps receiving traffic. Off by default: any cached probe error blocks the connection for the cache TTL (5 minutes).",
descriptionI18nKey: "featureFlagCredentialGateAuthDiagnosesOnlyDescription",
category: "health",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "CONTENT_POLICY_REFUSAL_COMBO_STOP_ENABLED",
label: "Combo: Stop on Content-Policy Refusal",
description:
"When a combo target answers with an exact content-policy refusal (code or type cyber_policy / content_policy_violation), stop the combo and return that refusal even when the remaining targets are different models, instead of advancing to the next target. Off by default: the refusal never cools the account or trips the provider breaker, and the combo advances to its next target exactly as before.",
descriptionI18nKey: "featureFlagContentPolicyRefusalComboStopEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
{
key: "COOLDOWN_RETRY_SKIP_REFUSALS_ENABLED",
label: "Cooldown Retry: Skip 403/410 Refusals",
description:
"When every account is cooling down and the cooldown that ends first was caused by a 403 (access refused) or a 410 (gone), return that refusal at once instead of waiting the cooldown out and re-sending the request; a renewing agentrouter connection-scoped quota 403 is then persisted with errorCode 429 so it stays retryable. Off by default: only a 401 skips the wait, and every refusal is persisted with the status it arrived with.",
descriptionI18nKey: "featureFlagCooldownRetrySkipRefusalsEnabledDescription",
category: "runtime",
defaultValue: "false",
type: "boolean",
requiresRestart: false,
warningLevel: "caution",
},
];