Freeway: self-hosted OpenAI-compatible gateway for free-tier providers

One endpoint in front of every free LLM tier, with automatic failover and no
provider key ever leaving the box.

Providers are data, not code: adding one is a JSON file plus an env var, never
a TypeScript change. Ships 12 providers pre-written; `verify-provider --apply`
derives the live model list from your own key.

- core: registry, bucketed quota meter (rolling + UTC-calendar windows),
  key-pool store, 6-step model router with diagnostic rejection reasons
- gateway: node:http, failover that commits to a stream only after 200 headers,
  SSE passthrough with usage sniffing, virtual keys with admin/inference roles
- context engine: sqlite sessions and a 5-rung refit ladder, so a 128k thread
  keeps working when it fails over to an 8k model
- cache: exact (sha256, safe-by-default) + opt-in semantic (flat Float32Array)
- guard: inbound limits, quota forecasting, limit calibration from upstream
  headers, key health, secret/PII/injection scanning
- dashboard: one self-contained HTML file, live SSE request log, token usage
- sdk: zero-dep typed client, drop-in for the OpenAI SDK

Zero runtime dependencies. 326 tests, 27-check smoke suite against a scripted
mock upstream so routing is testable without spending quota.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

providers/cloudflare.json

@@ -0,0 +1,89 @@
+{
+  "$schema": "./_schema.json",
+  "id": "cloudflare",
+  "label": "Cloudflare Workers AI",
+  "docs": "https://developers.cloudflare.com/workers-ai/models/",
+  "console": "https://dash.cloudflare.com/profile/api-tokens",
+  "enabled": true,
+  "priority": 20,
+  "adapter": "openai",
+  "baseUrl": "https://api.cloudflare.com/client/v4/accounts/${CLOUDFLARE_ACCOUNT_ID}/ai/v1",
+  "auth": {
+    "type": "bearer",
+    "envKeys": ["CLOUDFLARE_API_TOKEN"]
+  },
+  "limits": {
+    "creditsPerDay": 10000,
+    "creditsPerRequest": 15
+  },
+  "limitsSource": "docs",
+  "verifiedOn": "2026-08-16",
+  "modelsEndpoint": null,
+  "modelsPath": "data[].id",
+  "dropParams": [],
+  "headers": {},
+  "notes": "Free tier is 10,000 neurons/day (confirmed on the pricing page), metered per token rather than per request. Per-model neuron rates come straight from Cloudflare's published table; note output tokens cost roughly 7-10x input, which is why rates are per-model rather than one flat number. limits.creditsPerRequest is only a fallback for models without published rates. Model ids contain '/', which is why provider/model pins split on the FIRST slash only. The OpenAI-compatible surface exposes /chat/completions and /embeddings but not /models, hence modelsEndpoint: null.",
+  "models": [
+    {
+      "id": "@cf/meta/llama-3.1-8b-instruct",
+      "label": "Llama 3.1 8B Instruct",
+      "alias": ["cheap"],
+      "context": 8192,
+      "caps": ["chat", "code"],
+      "priority": 30,
+      "credits": { "perMTokIn": 2457, "perMTokOut": 18252 }
+    },
+    {
+      "id": "@cf/meta/llama-3.3-70b-instruct-fp8-fast",
+      "label": "Llama 3.3 70B Instruct (fp8 fast)",
+      "alias": ["big"],
+      "context": 24000,
+      "caps": ["chat", "tools", "code"],
+      "priority": 20,
+      "credits": { "perMTokIn": 26668, "perMTokOut": 204805 }
+    },
+    {
+      "id": "@cf/meta/llama-3.2-3b-instruct",
+      "label": "Llama 3.2 3B Instruct",
+      "alias": [],
+      "context": 8192,
+      "caps": ["chat"],
+      "priority": 40,
+      "credits": { "perMTokIn": 2457, "perMTokOut": 18252 }
+    },
+    {
+      "id": "@cf/mistralai/mistral-small-3.1-24b-instruct",
+      "label": "Mistral Small 3.1 24B",
+      "alias": [],
+      "context": 128000,
+      "caps": ["chat", "tools", "long-context"],
+      "priority": 25
+    },
+    {
+      "id": "@cf/zhipu/glm-4.7-flash",
+      "label": "GLM 4.7 Flash",
+      "alias": [],
+      "context": 131072,
+      "caps": ["chat", "tools", "reasoning", "long-context"],
+      "priority": 25
+    },
+    {
+      "id": "@cf/baai/bge-large-en-v1.5",
+      "label": "BGE Large EN v1.5",
+      "alias": ["embed-large"],
+      "context": 512,
+      "caps": ["embed"],
+      "priority": 20,
+      "credits": { "perMTokIn": 18582 }
+    },
+    {
+      "id": "@cf/baai/bge-small-en-v1.5",
+      "label": "BGE Small EN v1.5",
+      "alias": [],
+      "context": 512,
+      "caps": ["embed"],
+      "priority": 30,
+      "credits": { "perMTokIn": 1841 }
+    }
+  ]
+}