@oh-my-pi/pi-catalog 18.2.6 → 18.2.8

package/src/compat/rules/providers/local.kdl

@@ -0,0 +1,110 @@
+// On-device inference workers exposed as catalog models.
+
+provider "local" {
+	default-model "lfm2.5-230m"
+	allow-unauthenticated #true
+
+	seed api="local-inference" base-url="local://inference" bundle="always" {
+		model "kokoro" name="Kokoro-82M" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "parakeet-tdt-0.6b-v3" name="Parakeet TDT 0.6B v3" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "whisper-base" name="Whisper Base" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "whisper-small" name="Whisper Small" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "whisper-large-v3-turbo" name="Whisper Large v3 Turbo" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "lfm2.5-230m" name="LFM2.5 230M" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "lfm2.5-350m" name="LFM2.5 350M" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "falcon-h1-90m" name="Falcon H1 Tiny 90M" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "qwen3-1.7b" name="Qwen3 1.7B" {
+			reasoning #true
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "llama3.2:3b" name="Llama 3.2 3B" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "gemma-3-1b" name="Gemma 3 1B" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "qwen2.5-1.5b" name="Qwen2.5 1.5B" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "lfm2-1.2b" name="LFM2 1.2B" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+
+	models "kokoro" {
+		kind "tts"
+	}
+	models "parakeet-tdt-0.6b-v3" "whisper-*" {
+		kind "stt"
+	}
+	models "lfm*" "falcon-*" "qwen*" "llama3.2:3b" "gemma-3-1b" {
+		kind "tiny"
+	}
+}

package/src/compat/rules/providers/openai-codex.kdl

@@ -3,6 +3,24 @@
 provider "openai-codex" {
 	default-model "gpt-5.5"
 	env "OPENAI_CODEX_OAUTH_TOKEN"
+	kind-apis {
+		image "openai-codex-responses"
+	}
+	web-search "codex"
+
+	seed api="openai-codex-responses" base-url="https://chatgpt.com/backend-api" bundle="always" {
+		model "gpt-image-1" name="GPT Image 1" {
+			reasoning #false
+			input "text" "image"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+
+	models "gpt-image-1" {
+		kind "image"
+	}
 
 	// Replaces the Codex provider's handwritten flex/priority pricing fallback.
 	service-tier-cost {

package/src/compat/rules/providers/openai.kdl

@@ -3,6 +3,11 @@
 provider "openai" {
 	default-model "gpt-5.5"
 	env "OPENAI_API_KEY"
+	kind-apis {
+		embedding "openai-embeddings"
+		image "openai-responses"
+		stt "openai-transcriptions"
+	}
 
 	// Daybreak models are approval-gated first-party Responses models that are
 	// not yet present in stencil.so. Seed the documented aliases and current
@@ -30,6 +35,62 @@ provider "openai" {
 			cost input=12.5 output=75 cache-read=1.25 cache-write=15.625
 			limits context=400000 max-tokens=128000
 		}
+
+		// Catalog cost fields are per 1M tokens: GPT transcribers map documented
+		// audio-input/output-token rates directly. Whisper's $0.006/min duration
+		// tariff has no token-cost representation, so it remains zero and the
+		// transcription response's provider-reported usage/cost is authoritative.
+		model "whisper-1" name="Whisper 1" api="openai-transcriptions" base-url="https://api.openai.com/v1" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "gpt-4o-transcribe" name="GPT-4o Transcribe" api="openai-transcriptions" base-url="https://api.openai.com/v1" {
+			reasoning #false
+			input "text"
+			cost input=2.5 output=10 cache-read=0 cache-write=0
+			limits context=16000 max-tokens=2000
+			supports-tools #false
+		}
+		model "gpt-4o-mini-transcribe" name="GPT-4o Mini Transcribe" api="openai-transcriptions" base-url="https://api.openai.com/v1" {
+			reasoning #false
+			input "text"
+			cost input=1.25 output=5 cache-read=0 cache-write=0
+			limits context=16000 max-tokens=2000
+			supports-tools #false
+		}
+		// OpenAI publishes embedding prices per million input tokens; embedding
+		// responses have no output-token charge.
+		model "text-embedding-3-small" name="Text Embedding 3 Small" api="openai-embeddings" base-url="https://api.openai.com/v1" {
+			reasoning #false
+			input "text"
+			cost input=0.02 output=0 cache-read=0 cache-write=0
+			limits context=8192
+			supports-tools #false
+		}
+		model "text-embedding-3-large" name="Text Embedding 3 Large" api="openai-embeddings" base-url="https://api.openai.com/v1" {
+			reasoning #false
+			input "text"
+			cost input=0.13 output=0 cache-read=0 cache-write=0
+			limits context=8192
+			supports-tools #false
+		}
+		model "text-embedding-ada-002" name="Text Embedding Ada 002" api="openai-embeddings" base-url="https://api.openai.com/v1" {
+			reasoning #false
+			input "text"
+			cost input=0.1 output=0 cache-read=0 cache-write=0
+			limits context=8192
+			supports-tools #false
+		}
+	}
+
+	models "whisper-1" "gpt-4o-transcribe" "gpt-4o-mini-transcribe" {
+		kind "stt"
+	}
+	models "text-embedding-3-small" "text-embedding-3-large" "text-embedding-ada-002" {
+		kind "embedding"
 	}
 
 	// Replaces the strict-mode provider whitelist entry.

package/src/compat/rules/providers/openrouter.kdl

@@ -4,6 +4,133 @@ provider "openrouter" {
 	default-model "openai/gpt-5.5"
 	env "OPENROUTER_API_KEY"
 	discovery label="OpenRouter" allow-unauthenticated=#true
+	kind-apis {
+		embedding "openai-embeddings"
+		image "openrouter-images"
+		rerank "openrouter-rerank"
+		video "openrouter-video"
+		tts "openai-speech"
+		stt "openai-transcriptions"
+	}
+	web-search "openrouter"
+
+	// TypeSafe's Jev answers only through the Decisions API (`/api/alpha/decisions`,
+	// System One wire shape); the default `/models` roster omits `text->decisions`
+	// rows, so the alias is authored here and discovery refreshes the family.
+	seed api="openrouter-decisions" base-url="https://openrouter.ai/api/alpha" bundle="always" {
+		model "~typesafe/jev-latest" name="TypeSafe: Jev Latest" {
+			reasoning #false
+			input "text"
+			cost input=0.042 output=0 cache-read=0 cache-write=0
+			limits context=32000 max-tokens=28800
+		}
+
+		// OpenRouter's STT roster mixes per-token and duration-based prices.
+		// Token-priced GPT rows map per-token rates to catalog per-million costs;
+		// duration-priced rows remain zero because ModelCost has no seconds axis.
+		model "openai/whisper-1" name="OpenAI: Whisper 1" api="openai-transcriptions" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "openai/whisper-large-v3" name="OpenAI: Whisper Large V3" api="openai-transcriptions" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "openai/gpt-4o-transcribe" name="OpenAI: GPT-4o Transcribe" api="openai-transcriptions" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=2.5 output=10 cache-read=0 cache-write=0
+			limits context=128000 max-tokens=115200
+			supports-tools #false
+		}
+		model "microsoft/mai-transcribe-1.5" name="Microsoft AI: MAI-Transcribe 1.5" api="openai-transcriptions" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "microsoft/mai-transcribe-2" name="Microsoft AI: MAI-Transcribe 2" api="openai-transcriptions" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+
+		// OpenRouter bills reranking per 1,000 searches, which has no catalog cost
+		// axis. Keep token costs at zero; the provider-reported response cost is authoritative.
+		model "cohere/rerank-v3.5" name="Cohere: Rerank v3.5" api="openrouter-rerank" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits context=4096 max-tokens=3686
+			supports-tools #false
+		}
+		// Bundled fallbacks keep the gateway useful offline; live
+		// `/embeddings/models` discovery refreshes this roster and its per-token
+		// pricing. Catalog token costs are per million input tokens.
+		model "openai/text-embedding-3-small" name="OpenAI: Text Embedding 3 Small" api="openai-embeddings" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0.02 output=0 cache-read=0 cache-write=0
+			limits context=8192
+			supports-tools #false
+		}
+		model "qwen/qwen3-embedding-8b" name="Qwen: Qwen3 Embedding 8B" api="openai-embeddings" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text"
+			cost input=0.01 output=0 cache-read=0 cache-write=0
+			limits context=32768
+			supports-tools #false
+		}
+		// OpenRouter bills generated video by output second and resolution/SKU,
+		// which ModelCost cannot represent. Keep token costs at zero; poll-reported
+		// cost is authoritative. Live `/videos/models` discovery refreshes the roster.
+		model "google/veo-3.1" name="Google: Veo 3.1" api="openrouter-video" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text" "image"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "minimax/hailuo-3" name="MiniMax: H3" api="openrouter-video" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text" "image"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "alibaba/wan-2.7" name="Alibaba: Wan 2.7" api="openrouter-video" base-url="https://openrouter.ai/api/v1" {
+			reasoning #false
+			input "text" "image"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+	models "~typesafe/*" "typesafe/*" {
+		kind "judge"
+	}
+	models "cohere/rerank-v3.5" {
+		kind "rerank"
+	}
+	models "openai/text-embedding-3-small" "qwen/qwen3-embedding-8b" {
+		kind "embedding"
+	}
+	models "google/veo-3.1" "minimax/hailuo-3" "alibaba/wan-2.7" {
+		kind "video"
+	}
+	models "openai/whisper-1" "openai/whisper-large-v3" "openai/gpt-4o-transcribe" \
+		"microsoft/mai-transcribe-1.5" "microsoft/mai-transcribe-2" {
+		kind "stt"
+	}
 
 	// Replaces the OpenRouter provider wire-model-id dispatch branch.
 	wire-model-id-mode "openrouter"
@@ -59,6 +186,11 @@ provider "openrouter" {
 		thinking-efforts "minimal" "low" "medium" "high"
 		thinking-requires-effort #true
 	}
+	// The proxy conservatively classifies this tool-capable text+image output row
+	// as chat; OpenRouter's dedicated image roster confirms the image transport.
+	models "google/gemini-3-pro-image" {
+		kind "image"
+	}
 	// residue: taxonomy ranks and exact globs do not isolate these models.
 	models "qwen/qwen3-coder" {
 		thinking-format "openrouter"

package/src/compat/rules/providers/typesafe.kdl

@@ -0,0 +1,21 @@
+// TypeSafe System One judgment models.
+
+provider "typesafe" {
+	default-model "jev-latest"
+	env "TYPESAFE_API_KEY"
+
+	seed api="typesafe" base-url="https://api.typesafe.ai" bundle="always" {
+		model "jev-latest" name="TypeSafe jev" {
+			reasoning #false
+			input "text"
+			// Jev bills input only; matches the OpenRouter route seed.
+			cost input=0.042 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+
+	models "*" {
+		kind "judge"
+	}
+}

package/src/compat/rules/providers/web.kdl

@@ -0,0 +1,153 @@
+// Pure search engines exposed through the web-search runner.
+
+provider "web" {
+	default-model "public"
+	allow-unauthenticated #true
+
+	seed api="web-search" base-url="web://search" bundle="always" {
+		model "parallel" name="Parallel" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "perplexity" name="Perplexity" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "zai" name="Z.AI" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "exa" name="Exa" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "tinyfish" name="TinyFish" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "jina" name="Jina" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "kagi" name="Kagi" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "tavily" name="Tavily" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "firecrawl" name="Firecrawl" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "brave" name="Brave" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "kimi" name="Kimi" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "synthetic" name="Synthetic" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "ollama" name="Ollama" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "searxng" name="SearXNG" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "startpage" name="Startpage" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "duckduckgo" name="DuckDuckGo" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "ecosia" name="Ecosia" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "google" name="Google" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "mojeek" name="Mojeek" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "public" name="Public Web" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+
+	models "*" {
+		kind "search"
+	}
+}

package/src/compat/rules/providers/xai-oauth.kdl

@@ -4,6 +4,11 @@ provider "xai-oauth" {
 	default-model "grok-4.6"
 	env "XAI_OAUTH_TOKEN" "XAI_API_KEY"
 	discovery label="xAI Grok OAuth (SuperGrok)" oauth-provider="xai-oauth"
+	kind-apis {
+		image "openai-images"
+		tts "xai-tts"
+	}
+	web-search "xai"
 
 	// Source of truth for the xai-oauth chat picker. Declaration order is headline order.
 	// Context windows from hermes-agent/agent/model_metadata.py:205-220
@@ -82,6 +87,27 @@ provider "xai-oauth" {
 			cost input=0 output=0 cache-read=0 cache-write=0
 			limits context=200000 max-tokens=200000
 		}
+		model "grok-tts" name="Grok TTS" api="xai-tts" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+		model "grok-imagine-image" name="Grok Imagine Image" api="openai-images" {
+			reasoning #false
+			input "text" "image"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+
+	models "grok-imagine-image" {
+		kind "image"
+	}
+	models "grok-tts" {
+		kind "tts"
 	}
 
 	// Replaces the provider-keyed prompt-cache session-header baseline.

package/src/compat/rules/providers/xai.kdl

@@ -3,6 +3,28 @@
 provider "xai" {
 	default-model "grok-4.6"
 	env "XAI_API_KEY"
+	kind-apis {
+		image "openai-images"
+		tts "xai-tts"
+	}
+	web-search "xai"
+
+	seed api="xai-tts" base-url="https://api.x.ai/v1" bundle="always" {
+		model "grok-tts" name="Grok TTS" {
+			reasoning #false
+			input "text"
+			cost input=0 output=0 cache-read=0 cache-write=0
+			limits
+			supports-tools #false
+		}
+	}
+
+	models "grok-imagine-image" {
+		kind "image"
+	}
+	models "grok-tts" {
+		kind "tts"
+	}
 
 	// Replaces the provider-keyed prompt-cache session-header baseline.
 	prompt-cache-session-header "x-grok-conv-id"

package/src/compat/rules/runtime/behavior.kdl

@@ -140,7 +140,8 @@ behavior {
 	exclude-models provider="xiaomi-token-plan-sgp" substring="-tts" substring="-asr"
 	// SuperGrok's roster interleaves media models the chat picker cannot
 	// serve; the dedicated tool surfaces route them instead.
-	exclude-models provider="xai-oauth" prefix="grok-imagine-" prefix="grok-stt-" prefix="grok-voice-"
+	exclude-models provider="xai-oauth" exact="grok-imagine-image-quality" exact="grok-imagine-image-2.0" \
+		exact="grok-imagine-video" exact="grok-imagine-video-1.5" prefix="grok-stt-" prefix="grok-voice-"
 	// Meta's /v1/models lists image generation and transcription SKUs beside
 	// the Muse Spark chat models.
 	exclude-models provider="meta" prefix="muse-image-" prefix="muse-voice-"

package/src/compat/rules/taxonomy/qwen.kdl

@@ -13,6 +13,8 @@ class "qwen" {
     revision prefix="qwen"
 
     override id="yolo-auto-yolo-flash-identity" provider="yolo-auto" model="yolo" class="qwen" revision="3.8" rationale="Yolo-Auto's paid yolo route is initially backed by the same Qwen3.8 Flash deployment as qwen3.8-flash; the opaque alias carries no lineage tokens" provenance="yolo-auto.com/models (2026-09)"
+    override id="prismml-bonsai-27b-qwen-3-6" glob="*bonsai-27b*" class="qwen" revision="3.6" rationale="Reviewed PrismML Bonsai and Ternary-Bonsai 27B lineage is Qwen3.6-27B; the wire basename may carry a prefix or GGUF quantization suffix" provenance="packages/coding-agent/src/config/model-discovery.ts; original reviewed alias e3aa6594e5"
+    override id="prismml-bonsai-2-27b-qwen-3-8" glob="*bonsai-2-27b*" class="qwen" revision="3.8" rationale="Reviewed PrismML Bonsai 2 27B lineage is Qwen3.8-27B; the wire basename may carry a prefix or GGUF quantization suffix" provenance="packages/coding-agent/src/config/model-discovery.ts"
     override id="kilo-qwq-32b-family" provider="kilo" model="qwq-32b" logical="qwen/qwq-32b" class="qwen" rationale="The reviewed Kilo QwQ deployment belongs to the Qwen family despite its opaque product spelling" provenance="fixtures/llm-oracle/catalog/models.normalized.json"
     override id="nanogpt-eva-qwen-2-5-family" provider="nanogpt" model="EVA-Qwen2.5-32B-v0.2" logical="EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2" class="qwen" rationale="The reviewed EVA deployment is a Qwen 2.5 derivative despite its opaque product namespace" provenance="fixtures/llm-oracle/catalog/models.normalized.json"
     override id="nanogpt-eva-qwen-2-5-72b-family" provider="nanogpt" model="EVA-Qwen2.5-72B-v0.2" logical="EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2" class="qwen" rationale="The reviewed EVA 72B deployment is a Qwen 2.5 derivative despite its opaque product namespace" provenance="fixtures/llm-oracle/catalog/models.normalized.json"