@askalf/dario 4.8.45 → 4.8.47

package/dist/analytics.js

@@ -40,6 +40,9 @@ export function billingBucketFromClaim(claim) {
 // Anthropic pricing (per 1M tokens, USD). Not authoritative — used for
 // rough burn-rate display in the /analytics summary.
 const PRICING = {
+    // Fable 5 official per-token pricing wasn't published at integration time
+    // (2026-06-09) — assumed at the current flagship (opus-4-8) rate. Display-only.
+    'claude-fable-5': { input: 5, output: 25, cacheRead: 0.5, cacheCreate: 6.25 },
     'claude-opus-4-8': { input: 5, output: 25, cacheRead: 0.5, cacheCreate: 6.25 },
     'claude-opus-4-7': { input: 5, output: 25, cacheRead: 0.5, cacheCreate: 6.25 },
     'claude-opus-4-6': { input: 15, output: 75, cacheRead: 1.5, cacheCreate: 18.75 },
@@ -47,7 +50,11 @@ const PRICING = {
     'claude-haiku-4-5': { input: 0.8, output: 4, cacheRead: 0.08, cacheCreate: 1 },
 };
 function estimateCost(record) {
-    const p = PRICING[record.model] ?? PRICING['claude-sonnet-4-6'];
+    // Strip a trailing context tag (`claude-fable-5[1m]`, `claude-opus-4-7[1m]`)
+    // before the lookup — the [1m] ids billed at the wrong family's rate before
+    // (fell through to the sonnet fallback).
+    const baseModel = record.model.replace(/\[[^\]]*\]$/, '');
+    const p = PRICING[baseModel] ?? PRICING['claude-sonnet-4-6'];
     return ((record.inputTokens * p.input) +
         (record.outputTokens * p.output) +
         (record.cacheReadTokens * p.cacheRead) +

package/dist/cc-template.d.ts

@@ -291,14 +291,20 @@ export declare function parseEffortSuffix(model: string): {
  *               accepted by all and still routes to the subscription pool
  *               (verified: representative-claim=five_hour on Opus + Sonnet).
  *               Set --effort=xhigh / DARIO_EFFORT=xhigh for Opus's extra tier.)
+ *               EXCEPT fable: real CC's print-mode default for fable is
+ *               'high' (live capture 2026-06-09, CC v2.1.170 — the same
+ *               capture got 'xhigh' on opus and 'high' on fable), so the
+ *               unset-flag default mirrors that per-family. An explicit
+ *               flag still pins.
  *   'low' / 'medium' / 'high' / 'xhigh' / 'max' → pin to that value
  *   'ultracode' → 'xhigh' (CC's ultracode mode; xhigh on the wire)
  *   'client' → extract from `clientBody.output_config.effort` (normalized
- *              for the wire); fall back to 'max' if absent/non-string
+ *              for the wire); fall back to the per-family default if
+ *              absent/non-string
  *
  * Exported for tests.
  */
-export declare function resolveEffort(flag: EffortValue | undefined, clientBody: Record<string, unknown>): string;
+export declare function resolveEffort(flag: EffortValue | undefined, clientBody: Record<string, unknown>, model?: string): string;
 /**
  * Returns true if the given model accepts `thinking: { type: "adaptive" }`.
  *
@@ -308,6 +314,9 @@ export declare function resolveEffort(flag: EffortValue | undefined, clientBody:
  *   claude-opus-4-7    ✓ accepts adaptive
  *   claude-opus-4-6    ✓ accepts adaptive
  *   claude-sonnet-4-6  ✓ accepts adaptive
+ *   claude-fable-5     ✓ accepts adaptive (2026-06-09 — Fable 5 is CC's new
+ *                        flagship and real CC sends `thinking:{type:"adaptive"}`
+ *                        on it, incl. the `claude-fable-5[1m]` long-context id)
  *   claude-opus-4-5    ✗ "adaptive thinking is not supported on this model"
  *   claude-sonnet-4-5  ✗ same
  *   claude-haiku-4-5   ✗ same (already gated separately by isHaiku)

package/dist/cc-template.js

@@ -985,22 +985,29 @@ function normalizeEffortForWire(effort) {
  *               accepted by all and still routes to the subscription pool
  *               (verified: representative-claim=five_hour on Opus + Sonnet).
  *               Set --effort=xhigh / DARIO_EFFORT=xhigh for Opus's extra tier.)
+ *               EXCEPT fable: real CC's print-mode default for fable is
+ *               'high' (live capture 2026-06-09, CC v2.1.170 — the same
+ *               capture got 'xhigh' on opus and 'high' on fable), so the
+ *               unset-flag default mirrors that per-family. An explicit
+ *               flag still pins.
  *   'low' / 'medium' / 'high' / 'xhigh' / 'max' → pin to that value
  *   'ultracode' → 'xhigh' (CC's ultracode mode; xhigh on the wire)
  *   'client' → extract from `clientBody.output_config.effort` (normalized
- *              for the wire); fall back to 'max' if absent/non-string
+ *              for the wire); fall back to the per-family default if
+ *              absent/non-string
  *
  * Exported for tests.
  */
-export function resolveEffort(flag, clientBody) {
+export function resolveEffort(flag, clientBody, model) {
+    const familyDefault = (model ?? '').toLowerCase().includes('fable') ? 'high' : 'max';
     if (flag === undefined)
-        return 'max';
+        return familyDefault;
     if (flag === 'client') {
         const clientOC = clientBody.output_config;
         const clientEffort = clientOC?.effort;
         if (typeof clientEffort === 'string' && clientEffort.length > 0)
             return normalizeEffortForWire(clientEffort);
-        return 'max';
+        return familyDefault;
     }
     return normalizeEffortForWire(flag);
 }
@@ -1013,6 +1020,9 @@ export function resolveEffort(flag, clientBody) {
  *   claude-opus-4-7    ✓ accepts adaptive
  *   claude-opus-4-6    ✓ accepts adaptive
  *   claude-sonnet-4-6  ✓ accepts adaptive
+ *   claude-fable-5     ✓ accepts adaptive (2026-06-09 — Fable 5 is CC's new
+ *                        flagship and real CC sends `thinking:{type:"adaptive"}`
+ *                        on it, incl. the `claude-fable-5[1m]` long-context id)
  *   claude-opus-4-5    ✗ "adaptive thinking is not supported on this model"
  *   claude-sonnet-4-5  ✗ same
  *   claude-haiku-4-5   ✗ same (already gated separately by isHaiku)
@@ -1029,26 +1039,30 @@ export function resolveEffort(flag, clientBody) {
  */
 export function supportsAdaptiveThinking(modelId) {
     const m = modelId.toLowerCase();
-    // Opus/Sonnet, major-minor form: opus-4-6+, sonnet-4-6+, opus-5-X, etc.
+    // Opus/Sonnet/Fable, major-minor form: opus-4-6+, sonnet-4-6+, opus-5-X,
+    // fable-5-X, etc. (Fable launched at 5 — there is no fable-4 line — so the
+    // shared "4-6+" threshold is correct for it by construction.)
     //
     // Digit groups are bounded to {1,2} so the dated-suffix pre-4.x line
     // (`claude-3-5-sonnet-20241022`, `claude-3-7-sonnet-20250219`) doesn't
     // accidentally match the date as `sonnet-2024-1022` and parse year as
     // major. Realistic Anthropic version numbers are 1-2 digits.
-    const mm = m.match(/(?:opus|sonnet)-(\d{1,2})-(\d{1,2})\b/);
+    const mm = m.match(/(?:opus|sonnet|fable)-(\d{1,2})-(\d{1,2})\b/);
     if (mm) {
         const major = Number(mm[1]);
         const minor = Number(mm[2]);
         if (major > 4)
-            return true; // any opus-5+ / sonnet-5+
+            return true; // any opus-5+ / sonnet-5+ / fable-5+
         if (major === 4 && minor >= 6)
             return true; // 4-6, 4-7, …
         return false; // 4-5 and older
     }
-    // Major-only form (e.g. `opus-5`, `opus-10`). The negative lookahead
-    // prevents matching the `5` in `opus-5-X` (handled above), and the
-    // {1,2} bound prevents matching long dated suffixes.
-    const majorOnly = m.match(/(?:opus|sonnet)-(\d{1,2})(?!\d|-)/);
+    // Major-only form (e.g. `opus-5`, `fable-5`, `opus-10`). The negative
+    // lookahead prevents matching the `5` in `opus-5-X` (handled above), and
+    // the {1,2} bound prevents matching long dated suffixes. A trailing
+    // context tag (`claude-fable-5[1m]`) is fine: `[` is neither digit nor
+    // hyphen, so the lookahead passes.
+    const majorOnly = m.match(/(?:opus|sonnet|fable)-(\d{1,2})(?!\d|-)/);
     if (majorOnly && Number(majorOnly[1]) >= 5)
         return true;
     return false;
@@ -1470,7 +1484,7 @@ export function buildCCRequest(clientBody, billingTag, cacheControl, identity, o
             }
         }
         if (!skip || !skip.has('output_config')) {
-            ccRequest.output_config = { effort: resolveEffort(opts.effort, clientBody) };
+            ccRequest.output_config = { effort: resolveEffort(opts.effort, clientBody, model) };
         }
     }
     ccRequest.stream = stream;

package/dist/cli.js

@@ -1159,8 +1159,9 @@ async function help() {
 
   Proxy options:
     --model=MODEL            Force a model for all requests
-                             Shortcuts: opus, sonnet, haiku
-                             Full IDs: claude-opus-4-6, claude-sonnet-4-6
+                             Shortcuts: fable, fable1m, opus, sonnet, haiku
+                             Full IDs: claude-fable-5, claude-opus-4-8,
+                             claude-sonnet-4-6 (append [1m] for 1M context)
                              Provider prefix: openai:gpt-4o, groq:llama-3.3-70b,
                              claude:opus, local:qwen-coder (forces backend)
                              Default: passthrough (client decides)
@@ -1450,7 +1451,7 @@ async function help() {
     curl http://localhost:3456/v1/messages \\
       -H "Content-Type: application/json" \\
       -H "anthropic-version: 2023-06-01" \\
-      -d '{"model":"claude-opus-4-6","max_tokens":1024,"messages":[{"role":"user","content":"Hello"}]}'
+      -d '{"model":"claude-fable-5","max_tokens":1024,"messages":[{"role":"user","content":"Hello"}]}'
 
   Your subscription handles the billing. No API key needed.
   Tokens auto-refresh in the background — set it and forget it.

package/dist/doctor.js

@@ -479,6 +479,7 @@ export async function runChecks(opts = {}) {
                 { family: 'haiku', model: 'claude-haiku-4-5' },
                 { family: 'sonnet', model: 'claude-sonnet-4-6' },
                 { family: 'opus', model: 'claude-opus-4-8' },
+                { family: 'fable', model: 'claude-fable-5' },
             ];
             const probe = async (model) => {
                 const res = await fetch(probeEndpoint, {

package/dist/pool.d.ts

@@ -78,15 +78,18 @@ export interface PoolStatus {
 /** Parse an Anthropic response's rate-limit headers into a snapshot. */
 export declare function parseRateLimits(headers: Headers): RateLimitSnapshot;
 /**
- * Extract the model family (`opus` / `sonnet` / `haiku`) from a request's
- * model id. Used to look up the per-model 7d bucket in
+ * Extract the model family (`opus` / `sonnet` / `haiku` / `fable`) from a
+ * request's model id. Used to look up the per-model 7d bucket in
  * `RateLimitSnapshot.perModel7d` during routing decisions. Returns null
  * for non-Claude models or model ids that don't carry a recognizable
  * family token (those requests just use the unified buckets).
  *
  * Generous on input shape: matches `claude-opus-4-7`, `opus`, `claude-3-7-sonnet-…`,
- * `claude-haiku-4-5`, anything containing the family token. Lowercase-normalized
- * so it pairs cleanly with `parseRateLimits`'s lowercase family keys.
+ * `claude-haiku-4-5`, `claude-fable-5[1m]`, anything containing the family token.
+ * Lowercase-normalized so it pairs cleanly with `parseRateLimits`'s lowercase
+ * family keys (the header parser is generic on `7d_<family>`, so a `7d_fable`
+ * bucket is captured automatically the moment Anthropic starts emitting it —
+ * this function is what lets routing USE it).
  */
 export declare function modelFamily(modelId: string | null | undefined): string | null;
 /**

package/dist/pool.js

@@ -94,15 +94,18 @@ export function parseRateLimits(headers) {
     };
 }
 /**
- * Extract the model family (`opus` / `sonnet` / `haiku`) from a request's
- * model id. Used to look up the per-model 7d bucket in
+ * Extract the model family (`opus` / `sonnet` / `haiku` / `fable`) from a
+ * request's model id. Used to look up the per-model 7d bucket in
  * `RateLimitSnapshot.perModel7d` during routing decisions. Returns null
  * for non-Claude models or model ids that don't carry a recognizable
  * family token (those requests just use the unified buckets).
  *
  * Generous on input shape: matches `claude-opus-4-7`, `opus`, `claude-3-7-sonnet-…`,
- * `claude-haiku-4-5`, anything containing the family token. Lowercase-normalized
- * so it pairs cleanly with `parseRateLimits`'s lowercase family keys.
+ * `claude-haiku-4-5`, `claude-fable-5[1m]`, anything containing the family token.
+ * Lowercase-normalized so it pairs cleanly with `parseRateLimits`'s lowercase
+ * family keys (the header parser is generic on `7d_<family>`, so a `7d_fable`
+ * bucket is captured automatically the moment Anthropic starts emitting it —
+ * this function is what lets routing USE it).
  */
 export function modelFamily(modelId) {
     if (!modelId)
@@ -114,6 +117,8 @@ export function modelFamily(modelId) {
         return 'sonnet';
     if (m.includes('haiku'))
         return 'haiku';
+    if (m.includes('fable'))
+        return 'fable';
     return null;
 }
 /**

package/dist/proxy.d.ts

@@ -17,6 +17,18 @@ export declare function parseProviderPrefix(model: string): {
     provider: 'openai' | 'claude';
     model: string;
 } | null;
+/**
+ * Fable-5 (2026-06) — real CC appends `fallback-credit-2026-06-01` to the
+ * beta set on fable requests ONLY (live captures 2026-06-09, CC v2.1.170:
+ * the fable request carries it, the opus request does not). Subscription
+ * traffic on fable WITHOUT it is soft-refused upstream — every request
+ * returns 200 with `stop_reason: "refusal"` and empty content, while the
+ * same request on opus/sonnet answers normally. Mirror CC exactly: append
+ * for the fable family, leave every other family's set untouched.
+ * Exported for tests.
+ */
+export declare const FABLE_FALLBACK_CREDIT_BETA = "fallback-credit-2026-06-01";
+export declare function betaForModel(base: string, model: string | null | undefined): string;
 export declare const ORCHESTRATION_TAG_NAMES: string[];
 /**
  * Build the regex list that actually strips orchestration tags.

package/dist/proxy.js

@@ -128,6 +128,8 @@ function loadClaudeIdentity() {
 }
 // Model shortcuts — users can pass short names
 const MODEL_ALIASES = {
+    'fable': 'claude-fable-5',
+    'fable1m': 'claude-fable-5[1m]',
     'opus': 'claude-opus-4-8',
     'opus47': 'claude-opus-4-7',
     'opus46': 'claude-opus-4-6',
@@ -188,6 +190,24 @@ const BILLABLE_BETA_PREFIXES = [
 function filterBillableBetas(betas) {
     return betas.split(',').map(b => b.trim()).filter(b => b.length > 0 && !BILLABLE_BETA_PREFIXES.some(p => b.startsWith(p))).join(',');
 }
+/**
+ * Fable-5 (2026-06) — real CC appends `fallback-credit-2026-06-01` to the
+ * beta set on fable requests ONLY (live captures 2026-06-09, CC v2.1.170:
+ * the fable request carries it, the opus request does not). Subscription
+ * traffic on fable WITHOUT it is soft-refused upstream — every request
+ * returns 200 with `stop_reason: "refusal"` and empty content, while the
+ * same request on opus/sonnet answers normally. Mirror CC exactly: append
+ * for the fable family, leave every other family's set untouched.
+ * Exported for tests.
+ */
+export const FABLE_FALLBACK_CREDIT_BETA = 'fallback-credit-2026-06-01';
+export function betaForModel(base, model) {
+    if (!model || !model.toLowerCase().includes('fable'))
+        return base;
+    if (base.split(',').includes(FABLE_FALLBACK_CREDIT_BETA))
+        return base;
+    return base ? `${base},${FABLE_FALLBACK_CREDIT_BETA}` : FABLE_FALLBACK_CREDIT_BETA;
+}
 // Orchestration tags injected by agents (Aider, Cursor, OpenCode, etc.)
 // that confuse Claude when passed through. Strip before forwarding.
 export const ORCHESTRATION_TAG_NAMES = [
@@ -363,7 +383,7 @@ function translateStreamChunk(line) {
     catch { }
     return null;
 }
-const OPENAI_MODELS_LIST = { object: 'list', data: ['claude-opus-4-8', 'claude-opus-4-7', 'claude-opus-4-6', 'claude-sonnet-4-6', 'claude-haiku-4-5'].map(id => ({ id, object: 'model', created: 1700000000, owned_by: 'anthropic' })) };
+const OPENAI_MODELS_LIST = { object: 'list', data: ['claude-fable-5', 'claude-fable-5[1m]', 'claude-opus-4-8', 'claude-opus-4-7', 'claude-opus-4-6', 'claude-sonnet-4-6', 'claude-haiku-4-5'].map(id => ({ id, object: 'model', created: 1700000000, owned_by: 'anthropic' })) };
 /**
  * Append a JSON-ND line to the proxy log file. No-op when stream is
  * null (logFile not configured). Errors are swallowed — log writes
@@ -1560,7 +1580,9 @@ export async function startProxy(opts = {}) {
                 // skip context-1m entirely (dario#36).
                 const acctKey = poolAccount?.alias ?? ACCOUNT_KEY_SINGLE;
                 const skipContext1m = context1mUnavailable.has(acctKey);
-                beta = skipContext1m ? betaWithoutContext1m : betaBase;
+                // Fable requires its fallback-credit beta or upstream soft-refuses
+                // every request (see betaForModel) — model-conditional, like real CC.
+                beta = betaForModel(skipContext1m ? betaWithoutContext1m : betaBase, requestModel);
                 if (clientBeta) {
                     const baseSet = new Set(beta.split(','));
                     const filtered = filterBillableBetas(clientBeta)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@askalf/dario",
-  "version": "4.8.45",
+  "version": "4.8.47",
   "description": "Use your Claude Pro/Max subscription in any tool — Cursor, Cline, Aider, the Agent SDK, your scripts — at subscription pricing, not per-token API bills. One local Anthropic + OpenAI-compatible endpoint.",
   "type": "module",
   "bin": {