@oh-my-pi/pi-catalog 18.15.0 → 18.17.0

package/CHANGELOG.md

@@ -2,6 +2,44 @@
 
 ## [Unreleased]
 
+## [18.17.0] - 2026-11-06
+
+### Added
+
+- Added `tool-name-prefix` compat axis for Responses hosts that reject certain tool names; Muse Code sets it to `_` ([#17435](https://github.com/can1357/oh-my-pi/issues/17435), [#17515](https://github.com/can1357/oh-my-pi/pull/17515) by [@harshav167](https://github.com/harshav167), continued in [#18053](https://github.com/can1357/oh-my-pi/pull/18053) by [@abilliontokens](https://github.com/abilliontokens))
+- Added `supports-message-threads` compat axis (`supportsMessageThreads`), enabled for Claude 4.5+ on first-party Anthropic
+- Added opt-in Responses WebSocket capability via OMP settings or SDK preferences, off by default ([#17354](https://github.com/can1357/oh-my-pi/pull/17354) by [@TatsukiMeng](https://github.com/TatsukiMeng))
+- Added Factory Droid Gemini 4 Argon on Global inference for accounts with the feature flag ([#17861](https://github.com/can1357/oh-my-pi/pull/17861) by [@will-bogusz](https://github.com/will-bogusz))
+- Added Factory Droid Muse Spark 1.3 (Global region, `muse_spark_1_3` flag) with low to xhigh effort ([#17862](https://github.com/can1357/oh-my-pi/pull/17862) by [@will-bogusz](https://github.com/will-bogusz))
+
+### Changed
+
+- Updated Factory Droid client version reporting; improved tool-input streaming for Claude via Azure Foundry
+
+### Fixed
+
+- Fixed newly published Azure models missing after catalog refresh ([#18173](https://github.com/can1357/oh-my-pi/issues/18173), [#18174](https://github.com/can1357/oh-my-pi/pull/18174) by [@dunkpi](https://github.com/dunkpi))
+- Fixed Factory Droid hiding models when the live-routing family default yields no usable upstream; explicit empty per-model routes still hide them ([#17862](https://github.com/can1357/oh-my-pi/pull/17862) by [@will-bogusz](https://github.com/will-bogusz))
+
+## [18.16.0] - 2026-11-02
+
+### Added
+
+- `/login mistral` now offers the Mistral Vibe browser sign-in beside `MISTRAL_API_KEY` ([#13875](https://github.com/can1357/oh-my-pi/pull/13875) by [@Gauthier-Huguenin](https://github.com/Gauthier-Huguenin), continued in [#17672](https://github.com/can1357/oh-my-pi/pull/17672) by [@abilliontokens](https://github.com/abilliontokens))
+- Added `revisionAtLeast` to `compat/revision` and exported `withOpenAICompatibleDiscoveryTimeout` from `discovery/openai-compatible`
+- Added the Factory Droid models droid 1.1.0 serves: GPT-6.1 Terra, GPT-6 Sol and Luna Fast Mode, GPT-6 Astra and GPT-6.1 Sol Ultra Fast (after the organization's high-cost opt-in), Nemotron 3.5 Super, Muse Spark Open, Muse Spark 1.2 and Beam 1.0, each listed once the account may use it ([#17766](https://github.com/can1357/oh-my-pi/pull/17766) by [@will-bogusz](https://github.com/will-bogusz))
+- Added Factory Droid's dated credit rates: the credit rate follows Factory's promotions, discounted from a promotion's start and back to the base rate at its end, using the same dated rate cards as dollar prices (currently GPT-5.6 Sol and Gemini 3.7/3.8 Flash) ([#16309](https://github.com/can1357/oh-my-pi/pull/16309) by [@will-bogusz](https://github.com/will-bogusz))
+
+### Changed
+
+- Factory Droid lists Opus 5.5 Fast, GPT-6.1 Sol and MiniMax M3 without the feature flags Factory retired, and no longer routes GPT-5.5 to Snowflake ([#17766](https://github.com/can1357/oh-my-pi/pull/17766) by [@will-bogusz](https://github.com/will-bogusz))
+- Factory Droid Opus 5.5 now falls back to Opus 5 when Anthropic refuses a request, as droid does, while your account can use Opus 5 ([#17772](https://github.com/can1357/oh-my-pi/pull/17772) by [@will-bogusz](https://github.com/will-bogusz))
+
+### Fixed
+
+- GPT-3.5 and GPT-4 (Turbo) resolve the `cl100k` tokenizer, so their context estimates count in cl100k_base rather than the o200k_base default for models without one ([#17607](https://github.com/can1357/oh-my-pi/pull/17607) by [@will-bogusz](https://github.com/will-bogusz))
+- Added Scaleway Generative APIs as a built-in provider with 12 serverless chat models (glm-5.3 default), API-key login, and live model discovery ([#16687](https://github.com/can1357/oh-my-pi/pull/16687) by [@reda-maizate](https://github.com/reda-maizate))
+
 ## [18.15.0] - 2026-11-01
 
 ### Added

package/dist/types/compat/auth-ids.d.ts

@@ -1,4 +1,4 @@
 /** Every provider with an `auth/<id>.kdl` policy. */
-export type AuthProviderId = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";
+export type AuthProviderId = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "devin-cli" | "devin-windsurf" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "scaleway" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";
 /** Providers whose policy declares a `login` flow (the `/login` roster). */
-export type LoginProviderId = "abliteration" | "aiand" | "alibaba-coding-plan" | "alibaba-token-plan" | "anthropic" | "baseten" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google-antigravity" | "google-gemini-cli" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "meta" | "minimax-code" | "minimax-code-cn" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";
+export type LoginProviderId = "abliteration" | "aiand" | "alibaba-coding-plan" | "alibaba-token-plan" | "anthropic" | "baseten" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "devin-cli" | "devin-windsurf" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google-antigravity" | "google-gemini-cli" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "meta" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "scaleway" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";

package/dist/types/compat/catalog-policy.d.ts

@@ -1,7 +1,9 @@
 import type { CatalogPolicyInputs } from "../types.js";
+import { type ResolveRoute } from "./resolve.js";
 /**
- * The `catalog` axis bag authored in the KDL cascade rules for a model,
- * memoized per distinct model. Frozen: the bag is shared across callers, so it
- * must stay read-only — extract the value you need rather than handing it on.
+ * The `catalog` axis bag authored in the KDL cascade rules for a model (on one
+ * upstream when `route` names it), memoized per distinct model. Frozen: the bag
+ * is shared across callers, so it must stay read-only — extract the value you
+ * need rather than handing it on.
  */
-export declare function resolveCatalogPolicy(model: CatalogPolicyInputs): Readonly<Record<string, unknown>>;
+export declare function resolveCatalogPolicy(model: CatalogPolicyInputs, route?: ResolveRoute): Readonly<Record<string, unknown>>;

package/dist/types/compat/factory-droid.d.ts

@@ -1,3 +1,4 @@
+import { type DatedCostCard } from "../pricing.js";
 import type { Api, ModelSpec } from "../types.js";
 import { type FactoryDroidRegion, type FactoryDroidWire } from "../wire/factory-droid.js";
 /** Native token limits for one inference region. */
@@ -5,11 +6,16 @@ export interface FactoryDroidLimits {
     contextWindow: number | null;
     maxTokens: number | null;
 }
-/** Standard Credits rates: `input` is the per-token weight; `output` and `cacheRead` multiply it. */
+/**
+ * Standard Credits rates: `input` is the per-token weight; `output` and
+ * `cacheRead` multiply it. `effectiveRates` are dated cards (ascending) that
+ * overwrite those legs from their date, as on `cost-patch`.
+ */
 export interface FactoryDroidCreditRates {
     input: number;
     output?: number;
     cacheRead?: number;
+    effectiveRates?: readonly DatedCostCard[];
 }
 /** Account gates on one registry model (the `entitlement` axis). */
 export interface FactoryDroidEntitlement {

package/dist/types/compat/provider-ids.d.ts

@@ -1,2 +1,2 @@
 /** Every chat-model provider with a catalog entry (`default-model` in its provider rules). */
-export type KnownProvider = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kilo" | "kimi-code" | "litellm" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "opencode-go" | "opencode-zen" | "openrouter" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stepfun" | "synthetic" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zenmux" | "zhipu-coding-plan";
+export type KnownProvider = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kilo" | "kimi-code" | "litellm" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "opencode-go" | "opencode-zen" | "openrouter" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "scaleway" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stepfun" | "synthetic" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zenmux" | "zhipu-coding-plan";

package/dist/types/compat/resolve.d.ts

@@ -21,7 +21,7 @@ export declare function resolveDiscoveryApi(spec: ModelSpec<Api>, providerType:
  * {@link resolveModelPolicy} without resolving the compat and thinking
  * policies, which dominate its cost on catalog-wide scans.
  */
-export declare function resolveCatalogAxes(spec: CatalogPolicyInputs): Record<string, unknown>;
+export declare function resolveCatalogAxes(spec: CatalogPolicyInputs, route?: ResolveRoute): Record<string, unknown>;
 /**
  * Resolves the full policy surface for one model spec: structured identity,
  * complete compat record, thinking metadata, and catalog-data corrections.

package/dist/types/compat/responses-websocket.d.ts

@@ -0,0 +1,6 @@
+import type { Model } from "../types.js";
+/**
+ * Capability is rule-owned; callers must opt in. Either explicit false vetoes
+ * the other preference, including a per-request true for a disabled model.
+ */
+export declare function resolveResponsesWebSocketPreference(model: Model, requestPreference?: boolean): boolean;

package/dist/types/compat/revision.d.ts

@@ -22,6 +22,8 @@ export declare function parseRevision(value: string): Revision | undefined;
 export declare function parseRevisionPrefix(value: string): Revision | undefined;
 /** Lexicographic triple comparison: negative, zero, or positive. */
 export declare function compareRevision(a: Revision, b: Revision): number;
+/** Whether a revision string parses and is at or above `floor`; `undefined` or unparsable is never. */
+export declare function revisionAtLeast(revision: string | undefined, floor: string): boolean;
 /** Renders a revision as its canonical `major.minor.patch` string. */
 export declare function formatRevision(revision: Revision): string;
 /** Comparison operator inside a `revision ">=2.5 <4"` cascade constraint. */

package/dist/types/compat/server-side-fallback.d.ts

@@ -1,7 +1,9 @@
-import type { Model } from "../types.js";
+import type { CatalogPolicyInputs } from "../types.js";
+import type { ResolveRoute } from "./resolve.js";
 /**
  * Ordered model ids for Anthropic's server-side refusal `fallbacks` chain,
- * authored on the `server-side-fallback-models` catalog axis. Empty when no
- * rule assigns a chain (the model or host does not support the feature).
+ * authored on the `server-side-fallback-models` catalog axis, on `route`'s
+ * upstream when it names one. Empty when no rule assigns a chain (the model,
+ * host or upstream does not support the feature).
  */
-export declare function serverSideFallbackModels(model: Model): readonly string[];
+export declare function serverSideFallbackModels(model: CatalogPolicyInputs, route?: ResolveRoute): readonly string[];

package/dist/types/compat/types.d.ts

@@ -714,6 +714,8 @@ export interface ResolvedAxes {
 export interface RequestPolicy {
     completionsReasoningMode?: "none" | "effort" | "opt-in" | "forced-on";
     completionsReasoningHistory?: "omit" | "preserved" | "interleaved";
+    /** Send `allow_missing_thinking` with the effort so tool-call turns may arrive without reasoning. */
+    completionsAllowMissingThinking?: boolean;
     anthropicThinking?: "adaptive" | "adaptive-summarized" | "budget-interleaved" | "budget-effort";
     /** Advertise `fine-grained-tool-streaming-2025-05-14` on requests that carry tools. */
     anthropicToolStreamingBeta?: boolean;
@@ -721,12 +723,18 @@ export interface RequestPolicy {
     openaiPlatformHeader?: string;
     responsesCacheRetention?: boolean;
     responsesVerbosity?: "low";
-    responsesServiceTier?: "priority";
+    responsesServiceTier?: "priority" | "ultrafast";
     responsesParallelToolCalls?: boolean;
     responsesSafetyIdentifier?: boolean;
     /** Default `tool_choice` to `auto` when the request carries tools and the caller picks none. */
     responsesToolChoiceAuto?: boolean;
     googleThinking?: "level" | "level-medium";
+    /** Send no `temperature`, `topP` or `topK` in `generationConfig`. */
+    googleOmitSampling?: boolean;
+    /** Send `generationConfig.maxOutputTokens` as the model's output maximum. */
+    googlePinMaxOutputTokens?: boolean;
+    /** Replay the model's own turns as the native Gemini content it streamed. */
+    googleNativeReplay?: boolean;
     /** The route locks its upstream for the session (`x-provider-routing-source: session_lock`). */
     routingSessionLock?: boolean;
 }

package/dist/types/discovery/devin-proto.d.ts

@@ -1297,6 +1297,21 @@ export interface DevinPlanInfo extends ProtoMessage {
     accountDisplayName: string;
 }
 export declare const DevinPlanInfoSchema: MessageCodec<DevinPlanInfo>;
+/** Devin message exa.seat_management_pb.ExchangePKCEAuthorizationCodeRequest. */
+export interface ExchangePKCEAuthorizationCodeRequest extends ProtoMessage {
+    authorizationCode: string;
+    codeVerifier: string;
+}
+export declare const ExchangePKCEAuthorizationCodeRequestSchema: MessageCodec<ExchangePKCEAuthorizationCodeRequest>;
+/** Devin message exa.seat_management_pb.ExchangePKCEAuthorizationCodeResponse. */
+export interface ExchangePKCEAuthorizationCodeResponse extends ProtoMessage {
+    apiKey: string;
+    name: string;
+    apiServerUrl: string;
+    devinWebappHost: string;
+    devinApiUrl: string;
+}
+export declare const ExchangePKCEAuthorizationCodeResponseSchema: MessageCodec<ExchangePKCEAuthorizationCodeResponse>;
 /** Devin message exa.codeium_common_pb.ExperimentConfig. */
 export interface ExperimentConfig extends ProtoMessage {
     experiments: ExperimentWithVariant[];

package/dist/types/discovery/openai-compatible.d.ts

@@ -10,6 +10,12 @@ import type { Api, FetchImpl, ModelSpec, Provider } from "../types.js";
  * (issue #8315). 10s matches the coding-agent's remote-discovery budget.
  */
 export declare const DEFAULT_OPENAI_COMPATIBLE_DISCOVERY_TIMEOUT_MS = 10000;
+/**
+ * Uses a cancellable timer rather than the native abort-timeout helper so
+ * successful fast discovery requests do not leave armed timeout signals for
+ * concurrent GC to trip over later.
+ */
+export declare function withOpenAICompatibleDiscoveryTimeout<T>(timeoutMs: number, run: (signal: AbortSignal) => Promise<T>): Promise<T>;
 /**
  * Minimal OpenAI-style model entry shape consumed by discovery.
  *

package/dist/types/model-manager.d.ts

@@ -110,6 +110,8 @@ export interface ModelResolutionResult<TApi extends Api = Api> {
     stale: boolean;
     source: ModelResolutionSource;
     updatedAt?: number;
+    /** Why the provider's endpoint fetcher threw this call; the models then come from cache, models.dev or bundled data. */
+    fetchError?: string;
 }
 /**
  * Stateful facade over provider model resolution.

package/dist/types/models.d.ts

@@ -11,6 +11,22 @@ export declare function getBundledProviders(): readonly KnownProvider[];
 export declare function getBundledModels(provider: GeneratedProvider): Model<Api>[];
 /** Resolve display semantics without mistaking an absent rate card for a free model. */
 export declare function getModelPricingStatus(model: Pick<Model<Api>, "cost" | "pricingStatus">): ModelPricingStatus;
+/** Factory Droid credit rate in force at a time, and the next dated change to it. */
+export interface FactoryDroidCreditRate {
+    rate: number;
+    /** The earliest later `credit-rates` card that changes `rate`: the rate it sets and when (Unix ms). */
+    next?: {
+        rate: number;
+        at: number;
+    };
+}
+/**
+ * Factory Droid Standard Credits rate at `timestamp` (Unix ms, default now):
+ * the `input` of the `credit-rates` card due then (the latest at or before
+ * it, as for cost cards), else the base rate. Undefined for models without a
+ * credit rate.
+ */
+export declare function getFactoryDroidCreditRate(model: Pick<Model, "factoryDroidCredits" | "factoryDroidCreditCards">, timestamp?: number): FactoryDroidCreditRate | undefined;
 /**
  * The base token card in force at `timestamp` (Unix ms, default now): the
  * latest due `timeBased` effective rate, else the row's own legs. Price

package/dist/types/pricing.d.ts

@@ -1,24 +1,84 @@
 import type { EffectiveTokenCost, ModelCost, TimeBasedCost, TokenCost } from "./types.js";
+/** Serialized time-based pricing schedule, as cached model rows carry it. */
+export declare const TimeBasedCostShape: import("@oh-my-pi/omptype").FluentType<{
+    effectiveRates?: ({
+        cacheRead: number;
+        cacheWrite: number;
+        input: number;
+        output: number;
+    } & {
+        effectiveFrom: number;
+        longContext?: ({
+            cacheRead: number;
+            cacheWrite: number;
+            input: number;
+            output: number;
+        } & {
+            inputThreshold: number;
+            inputThresholdInclusive?: boolean | undefined;
+        }) | undefined;
+    })[] | undefined;
+    offPeakMultiplier: number;
+    peakWindows: {
+        endMinute: number;
+        startMinute: number;
+        weekdays: number[];
+    }[];
+}, {
+    effectiveRates?: ({
+        cacheRead: number;
+        cacheWrite: number;
+        input: number;
+        output: number;
+    } & {
+        effectiveFrom: number;
+        longContext?: ({
+            cacheRead: number;
+            cacheWrite: number;
+            input: number;
+            output: number;
+        } & {
+            inputThreshold: number;
+            inputThresholdInclusive?: boolean | undefined;
+        }) | undefined;
+    })[] | undefined;
+    offPeakMultiplier: number;
+    peakWindows: {
+        endMinute: number;
+        startMinute: number;
+        weekdays: number[];
+    }[];
+}>;
 /** Validate serialized schedules before admitting cached model rows. */
 export declare function isTimeBasedCost(value: unknown): value is TimeBasedCost;
 /** KDL object children are named records; normalize them once while constructing the model. */
 export declare function materializeTimeBasedCost(value: unknown): TimeBasedCost;
+/**
+ * The dated card due at `timestamp`: the one with the latest `effectiveFrom`
+ * at or before it, in any order, or undefined before the first card.
+ */
+export declare function dueDatedCard<T extends {
+    effectiveFrom: number;
+}>(cards: readonly T[], timestamp: number): T | undefined;
 /**
  * The latest `effectiveRates` card of a materialized schedule due at
  * `timestamp`, or undefined before the first one (or without a schedule).
  */
 export declare function dueEffectiveRate(schedule: TimeBasedCost | undefined, timestamp: number): EffectiveTokenCost | undefined;
-/** A `cost-patch`/`cost-fallback` dated card: the legs it overwrites from `effectiveFrom` (Unix ms). */
+/**
+ * A dated `effective-rates` card of a `cost-patch`, `cost-fallback` or
+ * `credit-rates` rule: the legs it overwrites from `effectiveFrom` (Unix ms).
+ */
 export interface DatedCostCard extends Partial<TokenCost> {
     effectiveFrom: number;
 }
 /**
- * Parse the compiled `effective-rates` block of a `cost-patch` or
- * `cost-fallback` payload into dated cards, ascending by date. Throws on a
- * malformed card, so a typo in a rule fails the build instead of silently
- * leaving the old price in force.
+ * Parse the compiled `effective-rates` block of a `cost-patch`,
+ * `cost-fallback` or Factory Droid `credit-rates` payload into dated cards,
+ * ascending by date. Throws on a malformed card, so a typo in a rule fails the
+ * build instead of silently leaving the old price in force.
  */
-export declare function parseDatedCostCards(costAxis: unknown, axis: "cost-patch" | "cost-fallback"): DatedCostCard[];
+export declare function parseDatedCostCards(costAxis: unknown, axis: "cost-patch" | "cost-fallback" | "credit-rates"): DatedCostCard[];
 /**
  * Earliest instant a rate card can name (the `Date` range minimum). A leading
  * card from here holds the price in force before the first dated change.

package/dist/types/provider-models/openai-compat.d.ts

@@ -740,3 +740,21 @@ export declare function singularityApiDevModelManagerOptions(config?: Singularit
  * reviewed KDL rules own the wire shape, limits patch, and effort ladder.
  */
 export declare function singularityApiTechModelManagerOptions(config?: SingularityApiModelManagerConfig): ModelManagerOptions<Api>;
+/**
+ * Scaleway discovery configuration: the API key plus optional base-URL and
+ * fetch overrides. Consumed by {@link scalewayModelManagerOptions}, and
+ * exported for extensions and tests that construct the manager directly.
+ */
+export interface ScalewayModelManagerConfig {
+    apiKey?: string;
+    baseUrl?: string;
+    fetch?: FetchImpl;
+}
+/**
+ * Scaleway model manager: plain OpenAI-compatible chat completions at
+ * `api.scaleway.ai/v1` (fr-par serverless Generative APIs). `/v1/models`
+ * returns bare ids, so the authored seed in `providers/scaleway.kdl` is the
+ * reference each live row merges names, modalities, prices, and limits from;
+ * a successful fetch is the complete roster and prunes retired rows.
+ */
+export declare function scalewayModelManagerOptions(config?: ScalewayModelManagerConfig): ModelManagerOptions<"openai-completions">;

package/dist/types/types.d.ts

@@ -1,5 +1,6 @@
 import type { ModelIdentity } from "./compat/types.js";
 import type { Effort } from "./effort.js";
+import type { DatedCostCard } from "./pricing.js";
 export type { FetchImpl } from "@oh-my-pi/pi-utils";
 export type { KnownProvider } from "./provider-models/descriptors.js";
 export type KnownApi = "openai-completions" | "openai-responses" | "openrouter" | "openai-codex-responses" | "azure-openai-responses" | "anthropic-messages" | "bedrock-converse-stream" | "google-generative-ai" | "google-gemini-cli" | "google-vertex" | "ollama-chat" | "cursor-agent" | "factory-droid-agent" | "gitlab-duo-agent" | "devin-agent" | "apple-foundation-models";
@@ -492,6 +493,14 @@ export interface AnthropicCompat {
      * Sonnet 4.6+, Fable/Mythos 5+ on supported hosts. Default: false.
      */
     supportsServerCompaction?: boolean;
+    /**
+     * Whether the model supports message threads (`message-threads-2026-08-12`):
+     * later turns continue a server-held conversation and send only new
+     * messages. Enabled on Claude 4.5+ for the first-party provider; the
+     * transport still requires the official endpoint unless the caller opts in.
+     * Default: false.
+     */
+    supportsMessageThreads?: boolean;
     /**
      * Whether the model is served by the first-party Anthropic provider (its
      * default route is the official API). Rule-owned on the provider; the
@@ -954,6 +963,14 @@ export interface ResolvedOpenAIResponsesCompat extends ResolvedOpenAISharedCompa
      * `PI_MUSE_STORE_RESPONSES`); stored runs retain prompts and outputs on the provider.
      */
     storeResponses: boolean;
+    /**
+     * Prefix the `openai-responses` handler puts on every function-tool name it
+     * declares or replays, and strips from returned calls through the request's
+     * wire-name table. Rule-owned: hosts that reject requests declaring certain
+     * exact tool names (Muse Code). Unset sends names unchanged. Other
+     * Responses-family handlers do not read it.
+     */
+    toolNamePrefix?: string;
     /**
      * Whether the host binds native history items (`encrypted_content`, item
      * ids) to the connection that issued them, so a new process must rebuild
@@ -1198,12 +1215,12 @@ export interface ModelCost extends TokenCost {
 /**
  * Exact local content tokenizer family for a model.
  *
- * Absent means no first-party local tokenizer is known and consumers retain
- * their estimate/default-tokenizer policy. The values name tokenizer
+ * Absent means no first-party local tokenizer is known and consumers count
+ * in o200k_base. The values name tokenizer
  * generations rather than providers: DeepSeek V3 through V4 share
  * `"deepseek-v3"`; Kimi K2 through K3 share `"kimi-k2"`.
  */
-export type ModelTokenizer = "claude-v3" | "claude-v47" | "claude-v5" | "claude-v5-sonnet" | "qwen3" | "deepseek-v3" | "kimi-k2" | "glm5" | "o200k";
+export type ModelTokenizer = "claude-v3" | "claude-v47" | "claude-v5" | "claude-v5-sonnet" | "qwen3" | "deepseek-v3" | "kimi-k2" | "glm5" | "cl100k" | "o200k";
 /** One account's discovered entitlements on a model; see {@link Model.accountAccess}. */
 export interface ModelAccountAccess {
     /**
@@ -1393,8 +1410,12 @@ export interface Model<TApi extends Api = Api> {
     factoryDroidApiProviders?: string[];
     /** Factory Droid: `configured_order` when live routing chose the rotation; absent means registry order. */
     factoryDroidRoutingSource?: "configured_order";
+    /** Factory Droid: server-side refusal fallbacks the account can run; absent when none are named. */
+    factoryDroidRefusalFallbacks?: string[];
     /** Factory Droid Standard Credits base rate (relative per-token weight, not dollars). */
     factoryDroidCredits?: number;
+    /** Factory Droid dated `credit-rates` cards, ascending; `getFactoryDroidCreditRate` picks the one due now. */
+    factoryDroidCreditCards?: readonly DatedCostCard[];
     /** Canonical organization resolved during Factory discovery. */
     factoryDroidOrgId?: string;
     cost: ModelCost;
@@ -1461,7 +1482,7 @@ export interface Model<TApi extends Api = Api> {
      * dispatch is redirected.
      */
     transport?: "pi-native";
-    /** Hint that websocket transport should be preferred when supported by the provider implementation. */
+    /** Standard Responses is opt-in and catalog-gated; either model/request false vetoes true. Codex keeps its independent policy. */
     preferWebsockets?: boolean;
     /** Codex Responses Lite transport: send the lite marker and carry instructions/tools as input items (mirrors codex-rs `use_responses_lite`). */
     useResponsesLite?: boolean;

package/dist/types/wire/codex.d.ts

@@ -46,7 +46,6 @@ export declare const OPENAI_HEADERS: {
     readonly ROUTING_HINT: "x-codex-routing-hint";
 };
 export declare const OPENAI_HEADER_VALUES: {
-    readonly BETA_RESPONSES: "responses=experimental";
     readonly BETA_RESPONSES_WEBSOCKETS_V2: "responses_websockets=2026-02-06";
     readonly REMOTE_COMPACTION_V2: "remote_compaction_v2";
     readonly ORIGINATOR_CODEX: "omp";

package/dist/types/wire/factory-droid.d.ts

@@ -3,7 +3,7 @@
  * native request corpus with `packages/ai/scripts/capture-factory-droid-native.ts`
  * against the matching CLI binary and make `factory-droid-native-parity.test.ts` pass.
  */
-export declare const FACTORY_DROID_CLIENT_VERSION = "0.237.0";
+export declare const FACTORY_DROID_CLIENT_VERSION = "1.1.0";
 /** Wire protocols the Factory proxy multiplexes; `api-routes provider="factory-droid"` picks one per model. */
 export declare const FACTORY_DROID_WIRES: readonly ["openai-completions", "openai-responses", "anthropic-messages", "google-generate"];
 /** Wire protocol used by the Factory proxy for each model. */

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@oh-my-pi/pi-catalog",
-	"version": "18.15.0",
+	"version": "18.17.0",
 	"description": "Model catalog for omp: bundled model database, provider discovery descriptors, model identity, classification, and equivalence",
 	"keywords": [
 		"ai",
@@ -78,7 +78,7 @@
 	},
 	"scripts": {
 		"check": "oxlint . && oxfmt --check --no-error-on-unmatched-pattern 'src/**/*.{ts,tsx}' '{test,bench,examples,scripts}/**/*.ts' '*.ts' && bun run check:types",
-		"check:types": "tsgo -p tsconfig.json --noEmit",
+		"check:types": "bun check --no-script -p tsconfig.json",
 		"lint": "oxlint .",
 		"test": "bun test --parallel",
 		"fix": "oxlint --fix --fix-suggestions . && bun run fmt",
@@ -88,12 +88,12 @@
 		"gen:proto": "bun scripts/generate-protocols.ts"
 	},
 	"dependencies": {
-		"@oh-my-pi/omptype": "18.15.0",
-		"@oh-my-pi/pi-utils": "18.15.0"
+		"@oh-my-pi/omptype": "18.17.0",
+		"@oh-my-pi/pi-utils": "18.17.0"
 	},
 	"devDependencies": {
 		"@bgotink/kdl": "0.4.0",
-		"@oh-my-pi/pi-ai": "18.15.0",
+		"@oh-my-pi/pi-ai": "18.17.0",
 		"@types/bun": "^1.3.14"
 	},
 	"engines": {

package/src/compat/auth-ids.ts

@@ -25,6 +25,8 @@ export type AuthProviderId =
 	| "deepinfra"
 	| "deepseek"
 	| "devin"
+	| "devin-cli"
+	| "devin-windsurf"
 	| "exa"
 	| "factory-droid"
 	| "firepass"
@@ -73,6 +75,7 @@ export type AuthProviderId =
 	| "qwen-portal"
 	| "sakana"
 	| "sambanova"
+	| "scaleway"
 	| "siliconflow"
 	| "siliconflow-cn"
 	| "singularityapi-dev"
@@ -122,6 +125,8 @@ export type LoginProviderId =
 	| "deepinfra"
 	| "deepseek"
 	| "devin"
+	| "devin-cli"
+	| "devin-windsurf"
 	| "exa"
 	| "factory-droid"
 	| "firepass"
@@ -144,6 +149,7 @@ export type LoginProviderId =
 	| "meta"
 	| "minimax-code"
 	| "minimax-code-cn"
+	| "mistral"
 	| "modular"
 	| "moonshot"
 	| "muse-code"
@@ -163,6 +169,7 @@ export type LoginProviderId =
 	| "qwen-portal"
 	| "sakana"
 	| "sambanova"
+	| "scaleway"
 	| "siliconflow"
 	| "siliconflow-cn"
 	| "singularityapi-dev"

package/src/compat/axes.ts

@@ -109,6 +109,7 @@ export const AXES: Readonly<Record<string, AxisDef>> = {
 		"preserved",
 		"interleaved",
 	]),
+	"completions-allow-missing-thinking": wire("completionsAllowMissingThinking", ["request"]),
 	"anthropic-thinking": wire("anthropicThinking", ["request"], "scalar", [
 		"adaptive",
 		"adaptive-summarized",
@@ -119,12 +120,15 @@ export const AXES: Readonly<Record<string, AxisDef>> = {
 	"openai-platform-header": wire("openaiPlatformHeader", ["request"]),
 	"responses-cache-retention": wire("responsesCacheRetention", ["request"]),
 	"responses-verbosity": wire("responsesVerbosity", ["request"], "scalar", ["low"]),
-	"responses-service-tier": wire("responsesServiceTier", ["request"], "scalar", ["priority"]),
+	"responses-service-tier": wire("responsesServiceTier", ["request"], "scalar", ["priority", "ultrafast"]),
 	"responses-parallel-tool-calls": wire("responsesParallelToolCalls", ["request"]),
 	"responses-safety-identifier": wire("responsesSafetyIdentifier", ["request"]),
 	"responses-tool-choice-auto": wire("responsesToolChoiceAuto", ["request"]),
 	"routing-session-lock": wire("routingSessionLock", ["request"]),
 	"google-thinking": wire("googleThinking", ["request"], "scalar", ["level", "level-medium"]),
+	"google-omit-sampling": wire("googleOmitSampling", ["request"]),
+	"google-pin-max-output-tokens": wire("googlePinMaxOutputTokens", ["request"]),
+	"google-native-replay": wire("googleNativeReplay", ["request"]),
 	// ── wire: OpenAI-compatible surfaces (chat completions + Responses) ──
 	"allows-synthetic-reasoning-content-for-tool-calls": wire("allowsSyntheticReasoningContentForToolCalls", OAI),
 	"always-send-max-tokens": wire("alwaysSendMaxTokens", OAI),
@@ -206,6 +210,7 @@ export const AXES: Readonly<Record<string, AxisDef>> = {
 	"supports-reasoning-params": wire("supportsReasoningParams", OAI),
 	"supports-reasoning-summary": wire("supportsReasoningSummary", ["openai-responses"]),
 	"store-responses": wire("storeResponses", ["openai-responses"]),
+	"tool-name-prefix": wire("toolNamePrefix", ["openai-responses"], "scalar", ["_"]),
 	"supports-store": wire("supportsStore", ["openai"]),
 	"supports-strict-mode": wire("supportsStrictMode", OAI),
 	"supports-tool-choice": wire("supportsToolChoice", OAI),
@@ -270,6 +275,7 @@ export const AXES: Readonly<Record<string, AxisDef>> = {
 	"supports-mid-conversation-tool-changes": wire("supportsMidConversationToolChanges", ["anthropic"]),
 	"supports-per-message-effort": wire("supportsPerMessageEffort", ["anthropic"]),
 	"supports-server-compaction": wire("supportsServerCompaction", ["anthropic", "bedrock"]),
+	"supports-message-threads": wire("supportsMessageThreads", ["anthropic"]),
 	"supports-between-tools-thinking": wire("supportsBetweenToolsThinking", ["anthropic"]),
 	"supports-thinking-binding-controls": wire("supportsThinkingBindingControls", ["anthropic", "bedrock"]),
 	"supports-turn-scoped-system": wire("supportsTurnScopedSystem", ["anthropic"]),
@@ -338,6 +344,7 @@ export const AXES: Readonly<Record<string, AxisDef>> = {
 	"thinking-unverified": { key: "unverified", set: "thinking", shape: "scalar" },
 
 	// ── catalog metadata ──
+	"supports-websockets": { key: "supportsWebsockets", set: "catalog", shape: "scalar", values: [true, false] },
 	"apply-patch-tool-type": {
 		key: "applyPatchToolType",
 		set: "catalog",
@@ -672,7 +679,9 @@ export const AXES: Readonly<Record<string, AxisDef>> = {
 	"region-limits-eu": { key: "regionLimitsEu", set: "catalog", shape: "object" },
 	/**
 	 * Subscription credit rates: `input` is the per-token credit weight shown
-	 * as the model's multiplier; `output` and `cache-read` multiply it.
+	 * as the model's multiplier; `output` and `cache-read` multiply it. An
+	 * `effective-rates` block of dated cards overwrites them from each card's
+	 * `effective-from`, as on `cost-patch`.
 	 */
 	"credit-rates": { key: "creditRates", set: "catalog", shape: "object" },
 	/**