deucebucket · deucebucket · Feb 19, 2026 · Feb 19, 2026 · Feb 19, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,6 +2,20 @@
 
 All notable changes to Library Manager will be documented in this file.
 
+## [0.9.0-beta.130] - 2026-02-18
+
+### Fixed
+
+- **Issue #160: Rate-limited batches no longer trigger false exhaustion** - Layer 4 processing
+  now distinguishes between "genuinely unidentifiable books" and "AI providers temporarily
+  unavailable." Rate-limited batches return a distinct signal (`-1`) and are not counted toward
+  the 3-strike exhaustion rule. When circuit breakers are open on AI providers, the worker waits
+  for recovery instead of marking books as "all processing layers exhausted." Previously, books
+  that were perfectly identifiable could be permanently marked as failed if providers were
+  rate-limited during processing.
+
+---
+
 ## [0.9.0-beta.129] - 2026-02-18
 
 ### Changed

diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@
 
 **Smart Audiobook Library Organizer with Multi-Source Metadata & AI Verification**
 
-[![Version](https://img.shields.io/badge/version-0.9.0--beta.129-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-0.9.0--beta.130-blue.svg)](CHANGELOG.md)
 [![Docker](https://img.shields.io/badge/docker-ghcr.io-blue.svg)](https://ghcr.io/deucebucket/library-manager)
 [![License](https://img.shields.io/badge/license-AGPL--3.0-blue.svg)](LICENSE)
 
@@ -16,6 +16,11 @@
 
 ## Recent Changes (stable)
 
+> **beta.130** - **Fix: Rate-Limited Batches No Longer Trigger False Exhaustion** (Issue #160)
+> - **Rate-limited batches skipped** - When AI providers are rate-limited, batches are no longer counted toward the 3-strike "all processing layers exhausted" rule
+> - **Circuit breaker awareness** - Layer 4 now waits for providers to recover instead of permanently marking identifiable books as failed
+> - **Distinct signal for rate limiting** - `process_queue` returns `-1` (rate-limited) vs `0` (genuinely empty) so the worker can react correctly
+
 > **beta.129** - **UI: Feedback Widget Moved to Nav Bar** (Issue #159)
 > - **Bug icon in nav bar** - Feedback/bug report button moved from floating bottom-right circle to a consistent bug icon in the top navigation bar
 > - **No more overlapping buttons** - Eliminates confusing dual floating buttons on the dashboard page

diff --git a/app.py b/app.py
@@ -11,7 +11,7 @@
 - Multi-provider AI (Gemini, OpenRouter, Ollama)
 """
 
-APP_VERSION = "0.9.0-beta.129"
+APP_VERSION = "0.9.0-beta.130"
 GITHUB_REPO = "deucebucket/library-manager"  # Your GitHub repo
 
 # Versioning Guide:
@@ -730,7 +730,7 @@
            try:
                with open(ERROR_REPORTS_PATH, 'r') as f:
                    reports = json.load(f)
            except:
                reports = []

        # Add new report (keep last 100 reports to avoid file bloat)
@@ -754,7 +754,7 @@
        try:
            with open(ERROR_REPORTS_PATH, 'r') as f:
                return json.load(f)
        except:
            return []
    return []

@@ -1709,7 +1709,7 @@
                    continue
                result = call_gemini(prompt, merged_config)
                if result:
                    logger.info(f"[PROVIDER CHAIN] Success with gemini")
                    return result

            elif provider == 'openrouter':
@@ -1718,13 +1718,13 @@
                    continue
                result = call_openrouter(prompt, merged_config)
                if result:
                    logger.info(f"[PROVIDER CHAIN] Success with openrouter")
                    return result

            elif provider == 'ollama':
                result = call_ollama(prompt, merged_config)
                if result:
                    logger.info(f"[PROVIDER CHAIN] Success with ollama")
                    return result

            else:
@@ -1826,7 +1826,7 @@
                        return result
                    elif result and result.get('transcript'):
                        # Got transcript but no match - still useful, return for potential AI fallback
                        logger.info(f"[AUDIO CHAIN] BookDB returned transcript only")
                        return result
                    elif result is None and attempt < max_retries - 1:
                        # Connection might be down, wait and retry
@@ -2158,11 +2158,11 @@
                device = "cuda"
                # int8 works on all CUDA devices including GTX 1080 (compute 6.1)
                # float16 only works on newer GPUs (compute 7.0+)
                logger.info(f"[WHISPER] Using CUDA GPU acceleration (10x faster)")
            else:
                logger.info(f"[WHISPER] Using CPU (no CUDA GPU detected)")
        except ImportError:
            logger.info(f"[WHISPER] Using CPU (ctranslate2 not available)")

        _whisper_model = WhisperModel(model_name, device=device, compute_type=compute_type)
        _whisper_model_name = model_name
@@ -2369,7 +2369,7 @@
        if sample_path and os.path.exists(sample_path):
            try:
                os.unlink(sample_path)
            except:
                pass

    return result

diff --git a/library_manager/pipeline/layer_ai_queue.py b/library_manager/pipeline/layer_ai_queue.py
@@ -100,7 +100,7 @@ def process_queue(
     allowed, calls_made, max_calls = check_rate_limit(config)
     if not allowed:
         logger.warning(f"Rate limit reached: {calls_made}/{max_calls} calls. Waiting...")
-        return 0, 0
+        return -1, 0  # Signal rate-limited (distinct from 0,0 = nothing to process)
 
     # Check if AI verification is enabled (before opening connection)
     if not config.get('enable_ai_verification', True):

diff --git a/library_manager/worker.py b/library_manager/worker.py
@@ -391,7 +391,33 @@ def process_all_queue(
         # At this point, we're trusting folder names as a last resort
         processed, fixed = process_queue(config, verification_layer=4)
 
+        # Issue #160: processed == -1 means rate-limited, NOT "nothing to process"
+        # Don't count rate-limited batches toward the 3-strike exhaustion rule
+        if processed == -1:
+            logger.info("Batch skipped due to rate limiting - not counting toward exhaustion")
+            _processing_status["current"] = "Rate limited, waiting for cooldown..."
+            _processing_status["last_activity"] = "Waiting for rate limit cooldown"
+            _processing_status["last_activity_time"] = time.time()
+            time.sleep(30)
+            continue
+
         if processed == 0:
+            # Check if AI providers are circuit-broken before counting as empty
+            # If providers are unavailable, this isn't a real "empty" result
+            ai_provider = config.get('ai_provider', 'gemini')
+            providers_to_check = [ai_provider]
+            if ai_provider != 'bookdb':
+                providers_to_check.append('bookdb')
+            any_circuit_open = any(is_circuit_open(p) for p in providers_to_check)
+
+            if any_circuit_open:
+                logger.info(f"AI providers circuit-broken ({', '.join(p for p in providers_to_check if is_circuit_open(p))}) - waiting for recovery, not counting toward exhaustion")
+                _processing_status["current"] = "AI provider cooling down, waiting..."
+                _processing_status["last_activity"] = "Waiting for circuit breaker recovery"
+                _processing_status["last_activity_time"] = time.time()
+                time.sleep(30)
+                continue
+
             conn = get_db()
             c = conn.cursor()
             c.execute('SELECT COUNT(*) as count FROM queue')