neoagent 3.0.1-beta.29 → 3.0.1-beta.30

package/docs/billing.md

@@ -61,6 +61,11 @@ Token limits are written directly to the user account when a subscription
 becomes active or is updated via webhook. The existing rate-limit enforcement
 in the runtime reads them with no additional configuration.
 
+Token limits gate interactive run starts. Background task executions are not
+blocked by the internal token admission check, so scheduled and integration
+tasks can still run while a user is over quota. Their token usage is still
+recorded and included in account and admin usage snapshots.
+
 ### Free plan
 
 Create a plan with **Price = 0** to serve as the default for new users. If a

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "neoagent",
-  "version": "3.0.1-beta.29",
+  "version": "3.0.1-beta.30",
   "description": "Self-hosted AI agent for long-running tasks, automation, messaging, device control, and local memory",
   "license": "AGPL-3.0-only",
   "main": "server/index.js",

package/server/public/.last_build_id

@@ -1 +1 @@
-8b097147c5ec8ac3bee082f22bd5d248
+400b929549327d2b7624182c525fe34e

package/server/public/flutter_bootstrap.js

@@ -37,6 +37,6 @@ _flutter.buildConfig = {"engineRevision":"a10d8ac38de835021c8d2f920dbf50a920ccc0
 
 _flutter.loader.load({
   serviceWorkerSettings: {
-    serviceWorkerVersion: "1039759445" /* Flutter's service worker is deprecated and will be removed in a future Flutter release. */
+    serviceWorkerVersion: "74968846" /* Flutter's service worker is deprecated and will be removed in a future Flutter release. */
   }
 });

package/server/public/main.dart.js

@@ -138719,7 +138719,7 @@ r===$&&A.b()
 p.push(A.iU(q,A.ju(!1,new A.Y(B.vr,A.cN(new A.cA(B.jU,new A.a94(r,q),q),q,q),q),!1,B.I,!0),q,q,0,0,0,q))}r=!1
 if(!s.ay)if(!s.ch){r=s.e
 r===$&&A.b()
-r=B.b.t("mr0924tg-6e8948b").length!==0&&r.b}if(r){r=s.d
+r=B.b.t("mr09cv28-5adc70e").length!==0&&r.b}if(r){r=s.d
 r===$&&A.b()
 r=r.aD&&!r.av?84:0
 s=s.e
@@ -145015,7 +145015,7 @@ $S:0}
 A.a0g.prototype={}
 A.U0.prototype={
 nm(a){var s=this
-if(B.b.t("mr0924tg-6e8948b").length===0||s.a!=null)return
+if(B.b.t("mr09cv28-5adc70e").length===0||s.a!=null)return
 s.Bg()
 s.a=A.m4(B.Td,new A.bgn(s))},
 Bg(){var s=0,r=A.l(t.H),q,p=2,o=[],n=this,m,l,k,j,i,h,g,f
@@ -145033,7 +145033,7 @@ if(!t.f.b(k)){s=1
 break}i=J.a2(k,"buildId")
 h=i==null?null:B.b.t(J.p(i))
 j=h==null?"":h
-if(J.bd(j)===0||J.f(j,"mr0924tg-6e8948b")){s=1
+if(J.bd(j)===0||J.f(j,"mr09cv28-5adc70e")){s=1
 break}n.b=!0
 n.E()
 p=2
@@ -145050,7 +145050,7 @@ case 2:return A.i(o.at(-1),r)}})
 return A.k($async$Bg,r)},
 vV(){var s=0,r=A.l(t.H),q,p=2,o=[],n=this,m,l,k,j,i,h,g,f,e,d,c,b,a,a0,a1
 var $async$vV=A.h(function(a2,a3){if(a2===1){o.push(a3)
-s=p}for(;;)switch(s){case 0:if(B.b.t("mr0924tg-6e8948b").length===0||n.c){s=1
+s=p}for(;;)switch(s){case 0:if(B.b.t("mr09cv28-5adc70e").length===0||n.c){s=1
 break}n.c=!0
 n.E()
 p=4

package/server/services/ai/loop/conversation_loop.js

@@ -571,7 +571,9 @@ async function runConversation(engine, userId, userMessage, options = {}, _model
   let deliverableWorkflow = null;
   const timelineService = app?.locals?.timelineService || null;
 
-  const { releaseReservation } = enforceRateLimits(userId);
+  const { releaseReservation } = enforceRateLimits(userId, {
+    bypass: options.bypassUserRateLimits === true,
+  });
 
   try {
   const historyWindow = Math.max(

package/server/services/ai/rate_limits.js

@@ -9,12 +9,7 @@ const _reservations = new Map();
 const DEFAULT_RATE_LIMIT_4H = 2_500_000;
 const DEFAULT_RATE_LIMIT_WEEKLY = 10_000_000;
 
-// Generous upper bound on a single run's token cost. Used to size the in-flight
-// reservation that keeps concurrent run starts from collectively bypassing the
-// per-user budget. It must NOT be the full limit — doing so would reserve the
-// entire budget for one run and reject any second concurrent run for the same
-// user (e.g. a scheduled task firing in the same minute as another).
-const DEFAULT_RUN_RESERVATION_TOKENS = 500_000;
+const MAX_RUN_RESERVATION_TOKENS = 100_000;
 
 const WINDOWS = {
   fourHour: {
@@ -142,7 +137,28 @@ function getRateLimitSnapshot(userId, { includeReservations = false } = {}) {
   };
 }
 
-function enforceRateLimits(userId) {
+function noopReleaseReservation() {}
+
+function calculateReservation(limits) {
+  const finiteLimits = [limits.fourHour, limits.weekly]
+    .filter((limit) => Number.isFinite(limit) && limit > 0);
+  if (finiteLimits.length === 0) return 1;
+  const minFiniteLimit = Math.min(...finiteLimits);
+  return Math.min(
+    MAX_RUN_RESERVATION_TOKENS,
+    Math.max(1, Math.floor(minFiniteLimit * 0.1)),
+  );
+}
+
+function enforceRateLimits(userId, options = {}) {
+  if (options.bypass === true) {
+    return {
+      snapshot: getRateLimitSnapshot(userId, { includeReservations: false }),
+      releaseReservation: noopReleaseReservation,
+      bypassed: true,
+    };
+  }
+
   const snapshot = getRateLimitSnapshot(userId, { includeReservations: true });
   if (snapshot.reached.fourHour) {
     throw new RateLimitExceededError('fourHour', snapshot);
@@ -150,23 +166,10 @@ function enforceRateLimits(userId) {
   if (snapshot.reached.weekly) {
     throw new RateLimitExceededError('weekly', snapshot);
   }
-  // Reserve a placeholder so concurrent starts see this run as in-flight.
-  // The reservation is released (or reconciled) when the run completes. Size it
-  // to a realistic per-run estimate — never the full limit — so concurrent runs
-  // for the same user are admitted normally and only a genuinely near-budget
-  // user gets throttled. Clamp so the reservation can never by itself exceed a
-  // (possibly small/custom) limit.
+  // Reserve a bounded placeholder so concurrent starts see this run as
+  // in-flight without letting one run consume the user's entire budget.
   const key = String(userId);
-  const limits = snapshot.limits;
-  const estimate = parsePositiveInteger(
-    process.env.NEOAGENT_RUN_RESERVATION_TOKENS,
-    DEFAULT_RUN_RESERVATION_TOKENS,
-  );
-  const cap = Math.min(
-    limits.fourHour == null ? Infinity : limits.fourHour,
-    limits.weekly == null ? Infinity : limits.weekly,
-  );
-  const reserve = Math.max(1, Math.min(estimate, Number.isFinite(cap) ? cap : estimate));
+  const reserve = calculateReservation(snapshot.limits);
   _reservations.set(key, (_reservations.get(key) || 0) + reserve);
   return { snapshot, releaseReservation: () => releaseReservation(userId, reserve) };
 }
@@ -185,6 +188,7 @@ function releaseReservation(userId, amount) {
 module.exports = {
   DEFAULT_RATE_LIMIT_4H,
   DEFAULT_RATE_LIMIT_WEEKLY,
+  MAX_RUN_RESERVATION_TOKENS,
   RateLimitExceededError,
   configuredDefaultLimits,
   enforceRateLimits,

package/server/services/tasks/runtime.js

@@ -567,11 +567,12 @@ class TaskRuntime {
         const finalPrompt = basePrompt + recoveryNote;
         const runOptions = {
           triggerType: task.trigger_type || 'schedule',
-          triggerSource: task.trigger_type || 'schedule',
+          triggerSource: executionMeta.triggerSource || task.trigger_type || 'schedule',
           agentId,
           app: this.app,
           conversationId,
           taskId,
+          bypassUserRateLimits: true,
           deliveryState,
           allowMultipleProactiveMessages: normalizedConfig.allowMultipleMessages === true || normalizedConfig.allow_multiple_messages === true,
           stageProactiveMessages: true,