pi-hypercharm-provider 1.3.0 → 1.3.2

package/.pi/fabric/mesh/state.json

@@ -1,95 +1,6 @@
 {
   "format": 1,
-  "entries": {
-    "topology/hosts/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85": {
-      "key": "topology/hosts/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85",
-      "value": {
-        "format": 1,
-        "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "rootId": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "identity": {
-          "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-          "name": "main",
-          "kind": "main",
-          "sessionId": "01a00efe-4e23-7b2b-8583-2e8493780899"
-        },
-        "startedAt": 1786958035702,
-        "updatedAt": 1786958850932,
-        "expiresAt": 1786958865932
-      },
-      "version": 196,
-      "updatedAt": 1786958850934,
-      "updatedBy": {
-        "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "name": "main",
-        "kind": "main",
-        "sessionId": "01a00efe-4e23-7b2b-8583-2e8493780899"
-      }
-    },
-    "sessions/01a00efe-4e23-7b2b-8583-2e8493780899": {
-      "key": "sessions/01a00efe-4e23-7b2b-8583-2e8493780899",
-      "value": {
-        "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "name": "Peer 01a00efe",
-        "kind": "peer",
-        "status": "running",
-        "runner": "pi",
-        "transport": "host",
-        "cwd": "/Users/monotykamary/VCS/working-remote/open-source/pi-hypercharm-provider",
-        "sessionId": "01a00efe-4e23-7b2b-8583-2e8493780899",
-        "model": "neuralwatt/kimi-k3",
-        "thinking": "max",
-        "startedAt": 1786958035702,
-        "updatedAt": 1786958850932,
-        "pendingMessages": false,
-        "local": false
-      },
-      "version": 196,
-      "updatedAt": 1786958850935,
-      "updatedBy": {
-        "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "name": "main",
-        "kind": "main",
-        "sessionId": "01a00efe-4e23-7b2b-8583-2e8493780899"
-      }
-    },
-    "topology/participants/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85": {
-      "key": "topology/participants/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85",
-      "value": {
-        "format": 1,
-        "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "kind": "root",
-        "rootId": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "ownerHostId": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "ownerIdentityId": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "name": "main",
-        "status": "running",
-        "runner": "pi",
-        "transport": "host",
-        "capabilities": [
-          "steer",
-          "followUp",
-          "fabric"
-        ],
-        "cwd": "/Users/monotykamary/VCS/working-remote/open-source/pi-hypercharm-provider",
-        "sessionId": "01a00efe-4e23-7b2b-8583-2e8493780899",
-        "model": "neuralwatt/kimi-k3",
-        "thinking": "max",
-        "startedAt": 1786958035702,
-        "updatedAt": 1786958850932,
-        "pendingMessages": false,
-        "controlProtocol": "v1"
-      },
-      "version": 196,
-      "updatedAt": 1786958850936,
-      "updatedBy": {
-        "id": "session:01a00efe-4e23-7b2b-8583-2e8493780899",
-        "name": "main",
-        "kind": "main",
-        "sessionId": "01a00efe-4e23-7b2b-8583-2e8493780899"
-      }
-    }
-  },
+  "entries": {},
   "versions": {
     "topology/hosts/e9758aa8c95305a8d60a8aa2c20fd53195591d479b2460e1273457793d3f1fec": 18,
     "sessions/019fe2c8-1dd2-7435-9fa1-fa664831df18": 17,
@@ -100,9 +11,9 @@
     "topology/hosts/3195ffccdcff6cb767344b2d1af6c5f3e1e3e4a9906a81129ffaf24352769d2a": 217,
     "sessions/01a00c57-b7aa-7ea3-b78a-11e22d3cdd86": 216,
     "topology/participants/3195ffccdcff6cb767344b2d1af6c5f3e1e3e4a9906a81129ffaf24352769d2a": 217,
-    "topology/hosts/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85": 196,
-    "sessions/01a00efe-4e23-7b2b-8583-2e8493780899": 196,
-    "topology/participants/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85": 196
+    "topology/hosts/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85": 388,
+    "sessions/01a00efe-4e23-7b2b-8583-2e8493780899": 387,
+    "topology/participants/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85": 388
   },
   "tombstoneOrder": [
     "sessions/019fe2c8-1dd2-7435-9fa1-fa664831df18",
@@ -113,6 +24,9 @@
     "topology/hosts/e0efb9d647e046c355b8bff12785e7805d774e7b89892edb38b4b29667ddc15f",
     "sessions/01a00c57-b7aa-7ea3-b78a-11e22d3cdd86",
     "topology/participants/3195ffccdcff6cb767344b2d1af6c5f3e1e3e4a9906a81129ffaf24352769d2a",
-    "topology/hosts/3195ffccdcff6cb767344b2d1af6c5f3e1e3e4a9906a81129ffaf24352769d2a"
+    "topology/hosts/3195ffccdcff6cb767344b2d1af6c5f3e1e3e4a9906a81129ffaf24352769d2a",
+    "sessions/01a00efe-4e23-7b2b-8583-2e8493780899",
+    "topology/participants/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85",
+    "topology/hosts/b98efd3cf26df4c69c6deae014475c4c0c080b1d0e5ef1cb0ebcc8dca7501c85"
   ]
 }

package/README.md

@@ -27,21 +27,21 @@ _Hyperoptimized coding models — DeepSeek, GLM, Kimi, Qwen, MiniMax, Gemma, GPT
 | Model | Type | Context | Max Tokens | Input Cost | Output Cost |
 |-------|------|---------|------------|------------|-------------|
 | DeepSeek V4 Flash | Text | 1.0M | 384K | $0.20 | $0.40 |
-| DeepSeek V4 Flash 0731 | Text | 1.0M | 384K | $0.20 | $0.40 |
+| DeepSeek V4 Flash 0731 | Text | 1.0M | 384K | $0.13 | $0.26 |
 | DeepSeek V4 Pro | Text | 1.0M | 384K | $2.40 | $4.80 |
 | DeepSeek V4 Pro 0813 | Text | 1.0M | 262K | $1.44 | $4.31 |
 | Gemma 4 26B A4B | Text | 256K | 26K | $0.12 | $0.42 |
-| GLM-5 | Text | 203K | 20K | $0.76 | $2.36 |
-| GLM-5.1 | Text | 203K | 3K | $1.52 | $4.79 |
+| GLM-5 | Text | 203K | 20K | $0.84 | $2.61 |
+| GLM-5.1 | Text | 203K | 3K | $1.36 | $4.40 |
 | GLM-5.2 | Text | 1.0M | 33K | $1.40 | $4.40 |
 | gpt-oss-120b | Text | 131K | 13K | $0.16 | $0.65 |
-| Kimi K2.5 | Text | 262K | 26K | $0.56 | $2.94 |
+| Kimi K2.5 | Text | 262K | 26K | $0.57 | $3.00 |
 | Kimi K2.6 | Text + Image | 262K | 26K | $0.95 | $4.00 |
 | Kimi K2.7 Code | Text + Image | 256K | 16K | $0.95 | $4.00 |
 | Kimi K3 | Text + Image | 1.0M | 131K | $3.00 | $15.00 |
-| Llama 3.3 70B Instruct | Text | 128K | 13K | $0.61 | $0.84 |
+| Llama 3.3 70B Instruct | Text | 128K | 13K | $0.64 | $0.77 |
 | Llama 4 Maverick 17B 128E Instruct FP8 | Text | 430K | 43K | $0.27 | $0.90 |
-| MiniMax M2.7 | Text | 262K | 7K | $0.43 | $1.62 |
+| MiniMax M2.7 | Text | 262K | 7K | $0.47 | $1.76 |
 | MiniMax M3 | Text + Image | 512K | 512K | $0.33 | $1.31 |
 | Qwen3 Coder 480B A35B Instruct INT4 Mixed AR | Text | 106K | 11K | $0.45 | $2.15 |
 | Qwen3 Next 80B A3B Instruct | Text | 262K | 26K | $0.12 | $1.14 |

package/index.ts

@@ -624,6 +624,9 @@ const CREDITS_MIN_INTERVAL_MS = 15_000;
 const ACCOUNT_FETCH_TIMEOUT_MS = 8_000;
 
 let statusAbort: AbortController | null = null;
+// Bumped on every session_start; async continuations compare against this to
+// drop work belonging to a replaced session (its ctx is stale and throws).
+let statusEpoch = 0;
 let lastCreditsFetchAt = 0;
 let creditsInFlight: Promise<void> | null = null;
 let metaFetched = false;
@@ -707,7 +710,30 @@ function currentProviderId(ctx: ExtensionContext): string | undefined {
 	}
 }
 
+function isStaleCtxError(err: unknown): boolean {
+	return err instanceof Error && err.message.includes("This extension ctx is stale");
+}
+
+// Render entry point: swallows the stale-ctx throw so a refresh racing a
+// session replacement (newSession/fork/switchSession/reload) can't crash pi.
 function updateStatus(ctx: ExtensionContext): void {
+	try {
+		renderStatus(ctx);
+	} catch (err) {
+		if (!isStaleCtxError(err)) throw err;
+	}
+}
+
+// Re-render once an async refresh lands, unless the session was replaced
+// meanwhile (epoch bump) — its ctx is stale and the render is obsolete anyway.
+function updateStatusAfter(promise: Promise<void>, ctx: ExtensionContext): void {
+	const epoch = statusEpoch;
+	void promise.then(() => {
+		if (epoch === statusEpoch) updateStatus(ctx);
+	});
+}
+
+function renderStatus(ctx: ExtensionContext): void {
 	const provider = currentProviderId(ctx);
 	const hiddenByOtherProvider =
 		statusConfig.hideOnOtherProvider && provider !== undefined && provider !== PROVIDER_ID;
@@ -791,7 +817,7 @@ function commitPending(ctx: ExtensionContext): void {
 	if (pendingSawOutOfCredits) {
 		pendingSawOutOfCredits = false;
 		// Re-fetch now so the balance reflects exhaustion immediately
-		void refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, true).then(() => updateStatus(ctx));
+		updateStatusAfter(refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, true), ctx);
 		if (!outOfCreditsNotified && ctx.hasUI) {
 			outOfCreditsNotified = true;
 			ctx.ui.notify("HyperCharm is out of Hypercredits — recharge at hyper.charm.land", "error");
@@ -854,7 +880,7 @@ async function handleStatusCommand(args: string, ctx: ExtensionContext): Promise
 		writeStatusConfig();
 		if (value !== "off" && key === "account") {
 			// Turning account on: make sure we have data to show
-			void refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, true).then(() => updateStatus(ctx));
+			updateStatusAfter(refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, true), ctx);
 			void refreshAccountMeta(cachedApiKey, statusAbort?.signal ?? undefined);
 		}
 		updateStatus(ctx);
@@ -929,7 +955,7 @@ async function configureStatusInteractive(ctx: ExtensionContext): Promise<void>
 			statusConfig.account = nextMode(statusConfig.account);
 			writeStatusConfig();
 			if (statusConfig.account !== "off") {
-				void refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, true).then(() => updateStatus(ctx));
+				updateStatusAfter(refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, true), ctx);
 				void refreshAccountMeta(cachedApiKey, statusAbort?.signal ?? undefined);
 			}
 			continue;
@@ -1003,11 +1029,13 @@ export default function (pi: ExtensionAPI) {
 	});
 
 	pi.on("session_start", async (_event, ctx) => {
+		const epoch = ++statusEpoch;
 		revalidateAbort?.abort();
 		revalidateAbort = new AbortController();
 		const signal = revalidateAbort.signal;
 		statusAbort?.abort();
 		statusAbort = new AbortController();
+		const statusSignal = statusAbort.signal;
 
 		loadStatusConfig();
 		resetStatusState();
@@ -1017,15 +1045,18 @@ export default function (pi: ExtensionAPI) {
 		pi.registerProvider(PROVIDER_ID, makeProviderConfig());
 
 		resolveApiKey(ctx.modelRegistry).then(() => {
+			// A session replacement while the key resolved invalidated the
+			// captured ctx (fast-resume, /new, /fork); nothing below may touch it.
+			if (epoch !== statusEpoch) return;
 			// Prefetch credits/team metadata only when a HyperCharm model is active
 			// (pi-neuralwatt also prefetches so the first turn ends with data, but
 			// gating here avoids API calls in sessions that never use the provider).
 			if (currentProviderId(ctx) === PROVIDER_ID) {
-				void refreshCredits(cachedApiKey, statusAbort!.signal, true).then(() => updateStatus(ctx));
-				void refreshAccountMeta(cachedApiKey, statusAbort!.signal).then(() => updateStatus(ctx));
+				updateStatusAfter(refreshCredits(cachedApiKey, statusSignal, true), ctx);
+				updateStatusAfter(refreshAccountMeta(cachedApiKey, statusSignal), ctx);
 			}
 			revalidateModels(cachedApiKey, embeddedModels, signal).then((freshBase) => {
-				if (freshBase && !signal.aborted) {
+				if (freshBase && epoch === statusEpoch && !signal.aborted) {
 					currentModels = buildModels(freshBase, customModels, patches);
 					pi.registerProvider(PROVIDER_ID, makeProviderConfig());
 				}
@@ -1037,7 +1068,7 @@ export default function (pi: ExtensionAPI) {
 		updateStatus(ctx);
 		const model: any = (event as any).model;
 		if (model?.provider === PROVIDER_ID && cachedApiKey) {
-			void refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, false).then(() => updateStatus(ctx));
+			updateStatusAfter(refreshCredits(cachedApiKey, statusAbort?.signal ?? undefined, false), ctx);
 			void refreshAccountMeta(cachedApiKey, statusAbort?.signal ?? undefined);
 		}
 	});

package/models.json

@@ -35,24 +35,24 @@
     "name": "DeepSeek V4 Flash 0731",
     "reasoning": true,
     "thinkingLevelMap": {
-      "off": null,
-      "minimal": null,
+      "off": "none",
+      "minimal": "minimal",
       "low": "low",
-      "medium": null,
+      "medium": "medium",
       "high": "high",
       "xhigh": "xhigh",
-      "max": null
+      "max": "max"
     },
     "input": [
       "text"
     ],
     "cost": {
-      "input": 0.2,
-      "output": 0.4,
-      "cacheRead": 0.04,
+      "input": 0.13,
+      "output": 0.26,
+      "cacheRead": 0.028,
       "cacheWrite": 0
     },
-    "contextWindow": 1000000,
+    "contextWindow": 1048576,
     "maxTokens": 384000,
     "compat": {
       "supportsStore": false,
@@ -140,10 +140,10 @@
       "text"
     ],
     "cost": {
-      "input": 0.124,
+      "input": 0.122,
       "output": 0.42,
       "cacheRead": 0,
-      "cacheWrite": 0.062
+      "cacheWrite": 0.061
     },
     "contextWindow": 256000,
     "maxTokens": 25600,
@@ -171,10 +171,10 @@
       "text"
     ],
     "cost": {
-      "input": 0.76,
-      "output": 2.356,
+      "input": 0.84,
+      "output": 2.612,
       "cacheRead": 0,
-      "cacheWrite": 0.38
+      "cacheWrite": 0.42
     },
     "contextWindow": 202752,
     "maxTokens": 20275,
@@ -190,11 +190,11 @@
     "name": "GLM-5.1",
     "reasoning": true,
     "thinkingLevelMap": {
-      "off": null,
+      "off": "off",
       "minimal": null,
       "low": null,
       "medium": null,
-      "high": "high",
+      "high": null,
       "xhigh": null,
       "max": "max"
     },
@@ -202,16 +202,16 @@
       "text"
     ],
     "cost": {
-      "input": 1.52432,
-      "output": 4.79072,
-      "cacheRead": 0.283088,
-      "cacheWrite": 0
+      "input": 1.36,
+      "output": 4.4,
+      "cacheRead": 0,
+      "cacheWrite": 0.68
     },
-    "contextWindow": 202800,
+    "contextWindow": 202750,
     "maxTokens": 3276,
     "compat": {
       "supportsStore": false,
-      "supportsReasoningEffort": true,
+      "supportsReasoningEffort": false,
       "thinkingFormat": "deepseek",
       "maxTokensField": "max_tokens"
     }
@@ -295,10 +295,10 @@
       "text"
     ],
     "cost": {
-      "input": 0.558,
-      "output": 2.94,
+      "input": 0.5744,
+      "output": 3.005,
       "cacheRead": 0,
-      "cacheWrite": 0.279
+      "cacheWrite": 0.2872
     },
     "contextWindow": 262144,
     "maxTokens": 26214,
@@ -329,8 +329,8 @@
     "cost": {
       "input": 0.95,
       "output": 4,
-      "cacheRead": 0,
-      "cacheWrite": 0.16
+      "cacheRead": 0.16,
+      "cacheWrite": 0
     },
     "contextWindow": 262000,
     "maxTokens": 26214,
@@ -384,8 +384,8 @@
     "cost": {
       "input": 3,
       "output": 15,
-      "cacheRead": 0,
-      "cacheWrite": 0.3
+      "cacheRead": 0.3,
+      "cacheWrite": 0
     },
     "contextWindow": 1048576,
     "maxTokens": 131072,
@@ -413,10 +413,10 @@
       "text"
     ],
     "cost": {
-      "input": 0.608,
-      "output": 0.838,
+      "input": 0.638,
+      "output": 0.768,
       "cacheRead": 0,
-      "cacheWrite": 0.304
+      "cacheWrite": 0.319
     },
     "contextWindow": 128000,
     "maxTokens": 12800,
@@ -466,10 +466,10 @@
       "text"
     ],
     "cost": {
-      "input": 0.426,
-      "output": 1.62,
+      "input": 0.47,
+      "output": 1.76,
       "cacheRead": 0,
-      "cacheWrite": 0.213
+      "cacheWrite": 0.235
     },
     "contextWindow": 262100,
     "maxTokens": 6553,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-hypercharm-provider",
-  "version": "1.3.0",
+  "version": "1.3.2",
   "description": "HyperCharm provider extension for pi - Access DeepSeek, GLM, Kimi, Qwen, MiniMax, Gemma, and GPT-OSS models through the Charm Hyper API",
   "type": "module",
   "main": "index.ts",

package/scripts/update-models.js

@@ -274,10 +274,9 @@ function transformModel(apiModel) {
 // ─── README generation ────────────────────────────────────────────────────────
 
 function formatCost(cost) {
-  if (cost === 0) return 'Free';
-  if (cost === null || cost === undefined) return '-';
-  if (cost < 0.01) return `$${cost.toFixed(4)}`;
-  return `$${cost.toFixed(2)}`;
+  if (cost === 0) return '—';
+  if (cost === null || cost === undefined) return '—';
+  return '$' + cost.toFixed(2);
 }
 
 function formatNumber(num) {
@@ -380,6 +379,31 @@ function updateDeprecatedModels(modelsJsonPath, newModels) {
   }
 }
 
+/**
+ * Grace-period deprecated models (deprecatedAt within TTL) with metadata stripped.
+ * Keeps the README table serving models that are delisted but still within their
+ * 14-day grace window.
+ */
+function withDeprecatedForReadme(models) {
+  const deprecatedPath = path.join(process.cwd(), 'deprecated-models.json');
+  let deprecated = {};
+  try {
+    const parsed = JSON.parse(fs.readFileSync(deprecatedPath, 'utf8'));
+    if (parsed && typeof parsed === 'object' && !Array.isArray(parsed)) deprecated = parsed;
+  } catch { /* no graveyard yet */ }
+  const now = Date.now();
+  const seen = new Set(models.map(m => m.id));
+  const extras = [];
+  for (const entry of Object.values(deprecated)) {
+    if (!entry || !entry.id || seen.has(entry.id)) continue;
+    const removedAt = Date.parse(entry.deprecatedAt || '');
+    if (Number.isNaN(removedAt) || now - removedAt > DEPRECATED_MODEL_TTL_MS) continue;
+    const m = { ...entry };
+    delete m.deprecatedAt;
+    extras.push(m);
+  }
+  return extras.length > 0 ? [...models, ...extras] : models;
+}
 async function main() {
   const apiKey = resolveApiKey();
   if (!apiKey) {
@@ -456,7 +480,7 @@ async function main() {
     }
 
     // Build merged models with patches for README
-    const readmeModels = buildModels(apiTransformed, customModels, patch);
+    const readmeModels = buildModels(withDeprecatedForReadme(apiTransformed), customModels, patch);
     readmeModels.sort((a, b) => a.name.localeCompare(b.name));
 
     // Update README