pi-umans-provider 1.0.2 → 1.0.3

package/.github/FUNDING.yml

@@ -0,0 +1,4 @@
+github: monotykamary
+ko_fi: monotykamary
+buy_me_a_coffee: monotykamary
+polar: monotykamary

package/README.md

@@ -30,7 +30,6 @@ _A [pi](https://github.com/earendil-works/pi-coding-agent) provider extension wi
 | Coder | Kimi K2.7-Code | 262K | ✅ | ✅ | 33K |
 | Flash | Qwen3.6-35B-A3B | 262K | ✅ | ✅ | 33K |
 | GLM 5.2 | GLM-5.2 | 406K | ❌ | ✅ | 131K |
-| GLM 5.2 NVFP4 (experimental, short test from Jun 29) | GLM-5.2 | 406K | ❌ | ✅ | 131K |
 | Kimi K2.7 Code | Kimi K2.7-Code | 262K | ✅ | ✅ | 33K |
 | Qwen3.6 35B A3B | Qwen3.6-35B-A3B | 262K | ✅ | ✅ | 33K |
 

package/index.ts

@@ -505,7 +505,7 @@ function resetIdleTimer(ctx: any): void {
   const signal = usageAbort?.signal;
   idleTimer = setTimeout(async () => {
     idleTimer = null;
-    // Skip while streaming — the optimistic value is in effect and agent_end
+    // Skip while streaming — the optimistic value is in effect and agent_settled
     // will reconcile with a real fetch.
     if (activeStreams > 0) return;
     if (signal?.aborted) return;
@@ -515,7 +515,7 @@ function resetIdleTimer(ctx: any): void {
   }, IDLE_POLL_MS);
 }
 
-// After agent_end the server still counts our session for a brief lag. Wait
+// After agent_settled the server still counts our session for a brief lag. Wait
 // it out, then take a clean idle baseline and re-arm the idle poll.
 function scheduleEndFetch(ctx: any): void {
   if (endFetchTimer) clearTimeout(endFetchTimer);
@@ -659,7 +659,7 @@ export default function (pi: ExtensionAPI) {
   // Optimistic +1: the moment our agent starts, we know the real concurrent
   // count is at least (others + 1). No API call — the server has a ~2s
   // registration lag, so fetching now would undercount us anyway. One span
-  // per prompt (agent_start → agent_end), so no flicker between tool turns.
+  // per prompt (agent_start → agent_settled), so no flicker between tool turns.
   pi.on("agent_start", async (_event, ctx) => {
     if (!isUmansModel(ctx)) return;
     if (idleTimer) {
@@ -672,7 +672,7 @@ export default function (pi: ExtensionAPI) {
 
   // Turn ended: drop our optimistic +1 and reconcile with the server after a
   // short settle so the server has dropped our session from its count.
-  pi.on("agent_end", async (_event, ctx) => {
+  pi.on("agent_settled", async (_event, ctx) => {
     if (!isUmansModel(ctx)) return;
     activeStreams = Math.max(0, activeStreams - 1);
     updateUsageStatus(ctx);

package/models.json

@@ -49,22 +49,6 @@
     "contextWindow": 405504,
     "maxTokens": 131071
   },
-  {
-    "id": "umans-glm-5.2-nvfp4",
-    "name": "Umans GLM 5.2 NVFP4 (experimental, short test from Jun 29)",
-    "reasoning": true,
-    "input": [
-      "text"
-    ],
-    "cost": {
-      "input": 0,
-      "output": 0,
-      "cacheRead": 0,
-      "cacheWrite": 0
-    },
-    "contextWindow": 405504,
-    "maxTokens": 131071
-  },
   {
     "id": "umans-kimi-k2.7",
     "name": "Umans Kimi K2.7 Code",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-umans-provider",
-  "version": "1.0.2",
+  "version": "1.0.3",
   "description": "UMANS provider extension for pi - Access Kimi K2.6, GLM 5.1, Qwen 3.6, and coding-optimized models through the UMANS API",
   "type": "module",
   "main": "index.ts",