fix(graphiti): address additional CodeRabbit review comments

- Fix FalkorDB default port from 6379 to 6380
- Update provider list comment to include Google AI
- Fix requirements.txt comment to mention both LLM and embeddings
- Add logging and warning for unimplemented tool calling in GoogleLLMClient
- Fix overly broad exception handling (catch only JSONDecodeError)
- Add Azure deployment name validation (LLM and embedding deployments)

CHANGELOG.md

@@ -1,3 +1,58 @@
+## 2.6.0 - Multi-Provider Graphiti Support & Platform Fixes
+
+### ✨ New Features
+
+- **Google AI Provider for Graphiti**: Full Google AI (Gemini) support for both LLM and embeddings in the Memory Layer
+  - Add GoogleLLMClient with gemini-2.0-flash default model
+  - Add GoogleEmbedder with text-embedding-004 default model
+  - UI integration for Google API key configuration with link to Google AI Studio
+- **Ollama LLM Provider in UI**: Add Ollama as an LLM provider option in Graphiti onboarding wizard
+  - Ollama runs locally and doesn't require an API key
+  - Configure Base URL instead of API key for local inference
+- **LLM Provider Selection UI**: Add provider selection dropdown to Graphiti setup wizard for flexible backend configuration
+- **Per-Project GitHub Configuration**: UI clarity improvements for per-project GitHub org/repo settings
+
+### 🛠️ Improvements
+
+- Enhanced Graphiti provider factory to support Google AI alongside existing providers
+- Updated env-handlers to properly populate graphitiProviderConfig from .env files
+- Improved type definitions with proper Graphiti provider config properties in AppSettings
+- Better API key loading when switching between providers in settings
+
+### 🐛 Bug Fixes
+
+- **node-pty Migration**: Replaced node-pty with @lydell/node-pty for prebuilt Windows binaries
+  - Updated all imports to use @lydell/node-pty directly
+  - Fixed "Cannot find module 'node-pty'" startup error
+- **GitHub Organization Support**: Fixed repository support for GitHub organization accounts
+  - Add defensive array validation for GitHub issues API response
+- **Asyncio Deprecation**: Fixed asyncio deprecation warning by using get_running_loop() instead of get_event_loop()
+- Applied ruff formatting and fixed import sorting (I001) in Google provider files
+
+### 🔧 Other Changes
+
+- Added google-generativeai dependency to requirements.txt
+- Updated provider validation to include Google/Groq/HuggingFace type assertions
+
+---
+
+## What's Changed
+
+- fix(graphiti): address CodeRabbit review comments by @adryserage in 679b8cd
+- fix(lint): sort imports in Google provider files by @adryserage in 1a38a06
+- feat(graphiti): add Google AI as LLM and embedding provider by @adryserage in fe69106
+- fix: GitHub organization repository support by @mojaray2k in 873cafa
+- feat(ui): add LLM provider selection to Graphiti onboarding by @adryserage in 4750869
+- fix(types): add missing AppSettings properties for Graphiti providers by @adryserage in 6680ed4
+- feat(ui): add Ollama as LLM provider option for Graphiti by @adryserage in a3eee92
+- fix(ui): address PR review feedback for Graphiti provider selection by @adryserage in b8a419a
+- fix(deps): update imports to use @lydell/node-pty directly by @adryserage in 2b61ebb
+- fix(deps): replace node-pty with @lydell/node-pty for prebuilt binaries by @adryserage in e1aee6a
+- fix: add UI clarity for per-project GitHub configuration by @mojaray2k in c9745b6
+- fix: add defensive array validation for GitHub issues API response by @mojaray2k in b3636a5
+
+---
+
 ## 2.5.5 - Enhanced Agent Reliability & Build Workflow
 
 ### ✨ New Features

CLAUDE.md

@@ -118,7 +118,7 @@ See [RELEASE.md](RELEASE.md) for detailed release process documentation.
 - **worktree.py** - Git worktree isolation for safe feature development
 - **memory.py** - File-based session memory (primary, always-available storage)
 - **graphiti_memory.py** - Optional graph-based cross-session memory with semantic search
-- **graphiti_providers.py** - Multi-provider factory for Graphiti (OpenAI, Anthropic, Azure, Ollama)
+- **graphiti_providers.py** - Multi-provider factory for Graphiti (OpenAI, Anthropic, Azure, Ollama, Google AI)
 - **graphiti_config.py** - Configuration and validation for Graphiti integration
 - **linear_updater.py** - Optional Linear integration for progress tracking
 
@@ -192,8 +192,8 @@ Dual-layer memory architecture:
 - Graph database with semantic search (FalkorDB)
 - Cross-session context retrieval
 - Multi-provider support (V2):
-  - LLM: OpenAI, Anthropic, Azure OpenAI, Ollama
-  - Embedders: OpenAI, Voyage AI, Azure OpenAI, Ollama
+  - LLM: OpenAI, Anthropic, Azure OpenAI, Ollama, Google AI (Gemini)
+  - Embedders: OpenAI, Voyage AI, Azure OpenAI, Ollama, Google AI
 
 Enable with: `GRAPHITI_ENABLED=true` + provider credentials. See `.env.example`.
 

README.md

@@ -248,12 +248,13 @@ The Memory Layer is a **hybrid RAG system** combining graph nodes with semantic
 **Architecture:**
 - **Backend**: FalkorDB (graph database) via Docker
 - **Library**: Graphiti for knowledge graph operations
-- **Providers**: OpenAI, Anthropic, Azure OpenAI, or Ollama (local/offline)
+- **Providers**: OpenAI, Anthropic, Azure OpenAI, Google AI, or Ollama (local/offline)
 
 | Setup | LLM | Embeddings | Notes |
 |-------|-----|------------|-------|
 | **OpenAI** | OpenAI | OpenAI | Simplest - single API key |
 | **Anthropic + Voyage** | Anthropic | Voyage AI | High quality |
+| **Google AI** | Gemini | Google | Single API key, fast inference |
 | **Ollama** | Ollama | Ollama | Fully offline |
 | **Azure** | Azure OpenAI | Azure OpenAI | Enterprise |
 
@@ -309,11 +310,12 @@ The `.auto-claude/` directory is gitignored and project-specific - you'll have o
 | `CLAUDE_CODE_OAUTH_TOKEN` | Yes | OAuth token from `claude setup-token` |
 | `AUTO_BUILD_MODEL` | No | Model override (default: claude-opus-4-5-20251101) |
 | `GRAPHITI_ENABLED` | Recommended | Set to `true` to enable Memory Layer |
-| `GRAPHITI_LLM_PROVIDER` | For Memory | LLM provider: openai, anthropic, azure_openai, ollama |
-| `GRAPHITI_EMBEDDER_PROVIDER` | For Memory | Embedder: openai, voyage, azure_openai, ollama |
+| `GRAPHITI_LLM_PROVIDER` | For Memory | LLM provider: openai, anthropic, azure_openai, ollama, google |
+| `GRAPHITI_EMBEDDER_PROVIDER` | For Memory | Embedder: openai, voyage, azure_openai, ollama, google |
 | `OPENAI_API_KEY` | For OpenAI | Required for OpenAI provider |
 | `ANTHROPIC_API_KEY` | For Anthropic | Required for Anthropic LLM |
 | `VOYAGE_API_KEY` | For Voyage | Required for Voyage embeddings |
+| `GOOGLE_API_KEY` | For Google | Required for Google AI (Gemini) provider |
 
 See `auto-claude/.env.example` for complete configuration options.
 

auto-claude-ui/src/main/ipc-handlers/env-handlers.ts

@@ -151,7 +151,7 @@ ${existingVars['ENABLE_FANCY_UI'] !== undefined ? `ENABLE_FANCY_UI=${existingVar
 
 # =============================================================================
 # GRAPHITI MEMORY INTEGRATION (OPTIONAL)
-# Multi-provider support: OpenAI, Anthropic, Azure OpenAI, Ollama, Voyage
+# Multi-provider support: OpenAI, Anthropic, Google AI, Azure OpenAI, Ollama, Voyage
 # =============================================================================
 ${existingVars['GRAPHITI_ENABLED'] ? `GRAPHITI_ENABLED=${existingVars['GRAPHITI_ENABLED']}` : '# GRAPHITI_ENABLED=false'}
 

auto-claude-ui/src/renderer/components/onboarding/GraphitiStep.tsx

@@ -106,7 +106,7 @@ export function GraphitiStep({ onNext, onBack, onSkip }: GraphitiStepProps) {
   const { settings, updateSettings } = useSettingsStore();
   const [config, setConfig] = useState<GraphitiConfig>({
     enabled: false,
-    falkorDbUri: 'bolt://localhost:6379',
+    falkorDbUri: 'bolt://localhost:6380',
     llmProvider: 'openai',
     embeddingProvider: 'openai',
     openaiApiKey: settings.globalOpenAIApiKey || '',
@@ -186,6 +186,12 @@ export function GraphitiStep({ onNext, onBack, onSkip }: GraphitiStepProps) {
     if (llmProvider === 'azure_openai' || embeddingProvider === 'azure_openai') {
       if (!config.azureOpenaiApiKey.trim()) return 'Azure OpenAI API key';
       if (!config.azureOpenaiBaseUrl.trim()) return 'Azure OpenAI Base URL';
+      if (llmProvider === 'azure_openai' && !config.azureOpenaiLlmDeployment.trim()) {
+        return 'Azure OpenAI LLM deployment name';
+      }
+      if (embeddingProvider === 'azure_openai' && !config.azureOpenaiEmbeddingDeployment.trim()) {
+        return 'Azure OpenAI embedding deployment name';
+      }
     }
     if (embeddingProvider === 'voyage') {
       if (!config.voyageApiKey.trim()) return 'Voyage API key';

auto-claude/.env.example

@@ -140,10 +140,10 @@
 # Choose which providers to use for LLM and embeddings.
 # Default is "openai" for both.
 
-# LLM provider: openai | anthropic | azure_openai | ollama
+# LLM provider: openai | anthropic | azure_openai | ollama | google
 # GRAPHITI_LLM_PROVIDER=openai
 
-# Embedder provider: openai | voyage | azure_openai | ollama
+# Embedder provider: openai | voyage | azure_openai | ollama | google
 # GRAPHITI_EMBEDDER_PROVIDER=openai
 
 # =============================================================================
@@ -191,6 +191,23 @@
 # Available: voyage-3 (1024 dim), voyage-3-lite (512 dim)
 # VOYAGE_EMBEDDING_MODEL=voyage-3
 
+# =============================================================================
+# GRAPHITI: Google AI Provider
+# =============================================================================
+# Use Google AI (Gemini) for both LLM and embeddings.
+# Get API key from: https://aistudio.google.com/apikey
+#
+# Required: GOOGLE_API_KEY
+
+# Google AI API Key
+# GOOGLE_API_KEY=AIzaSyxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
+
+# Google LLM Model (default: gemini-2.0-flash)
+# GOOGLE_LLM_MODEL=gemini-2.0-flash
+
+# Google Embedding Model (default: text-embedding-004)
+# GOOGLE_EMBEDDING_MODEL=text-embedding-004
+
 # =============================================================================
 # GRAPHITI: Azure OpenAI Provider
 # =============================================================================
@@ -331,3 +348,9 @@
 # AZURE_OPENAI_BASE_URL=https://your-resource.openai.azure.com/...
 # AZURE_OPENAI_LLM_DEPLOYMENT=gpt-5
 # AZURE_OPENAI_EMBEDDING_DEPLOYMENT=text-embedding-3-small
+#
+# --- Example 5: Google AI (Gemini) ---
+# GRAPHITI_ENABLED=true
+# GRAPHITI_LLM_PROVIDER=google
+# GRAPHITI_EMBEDDER_PROVIDER=google
+# GOOGLE_API_KEY=AIzaSyxxxxxxxx

auto-claude/integrations/graphiti/providers_pkg/llm_providers/google_llm.py

@@ -6,10 +6,13 @@ Google Gemini LLM client implementation for Graphiti.
 Uses the google-generativeai SDK.
 """
 
+import logging
 from typing import TYPE_CHECKING, Any
 
 from ..exceptions import ProviderError, ProviderNotInstalled
 
+logger = logging.getLogger(__name__)
+
 if TYPE_CHECKING:
     from graphiti_config import GraphitiConfig
 
@@ -116,8 +119,11 @@ class GoogleLLMClient:
             try:
                 data = json.loads(response.text)
                 return response_model(**data)
-            except (json.JSONDecodeError, Exception):
-                # If parsing fails, return raw text
+            except json.JSONDecodeError:
+                # If JSON parsing fails, return raw text
+                logger.warning(
+                    "Failed to parse JSON response from Google AI, returning raw text"
+                )
                 return response.text
         else:
             response = await loop.run_in_executor(
@@ -135,16 +141,22 @@ class GoogleLLMClient:
         """
         Generate a response with tool calling support.
 
+        Note: Tool calling is not yet implemented for Google AI provider.
+        This method will log a warning and fall back to regular generation.
+
         Args:
             messages: List of message dicts
             tools: List of tool definitions
             **kwargs: Additional arguments
 
         Returns:
-            Generated response with potential tool calls
+            Generated response (without tool calls)
         """
-        # For now, fall back to regular generation
-        # Tool calling can be added later if needed
+        if tools:
+            logger.warning(
+                "Google AI provider does not yet support tool calling. "
+                "Tools will be ignored and regular generation will be used."
+            )
         return await self.generate_response(messages, **kwargs)
 
 

auto-claude/requirements.txt

@@ -5,5 +5,5 @@ python-dotenv>=1.0.0
 # Memory Integration (highly recommended) but can be disabled by commenting out the line below
 graphiti-core[falkordb]>=0.5.0
 
-# Google AI embeddings (optional - for Gemini embeddings)
+# Google AI (optional - for Gemini LLM and embeddings)
 google-generativeai>=0.8.0