Claude/add api compatible ollama 011 c urd mqq1x q jv scwxg gu1q

.env.template

@@ -1,6 +1,53 @@
 SYSTEM_PROMPT_PATH=./prompts/research_agent_prompt.txt
+
+# =============================================================================
+# LLM Provider Configuration
+# =============================================================================
+# Choose your provider: ollama, lmstudio, openai, openai_compatible
+LLM_PROVIDER=ollama
+EMBEDDING_PROVIDER=ollama
+
+# Model Configuration
+CHAT_MODEL=gpt-oss:120b-cloud
+EMBED_MODEL=nomic-embed-text
+TEMPERATURE=0.7
+
+# -----------------------------------------------------------------------------
+# Ollama Configuration
+# -----------------------------------------------------------------------------
+# For local Ollama: http://localhost:11434
+# For remote/API Ollama: http://your-ollama-server:11434
+OLLAMA_BASE_URL=http://localhost:11434
+OLLAMA_API_KEY=  # Optional, leave empty if not needed
+
+# -----------------------------------------------------------------------------
+# LM Studio Configuration
+# -----------------------------------------------------------------------------
+# Default LM Studio local server
+LMSTUDIO_BASE_URL=http://localhost:1234/v1
+LMSTUDIO_API_KEY=  # Optional, leave empty if not needed
+
+# -----------------------------------------------------------------------------
+# OpenAI Configuration
+# -----------------------------------------------------------------------------
+OPENAI_API_KEY=<your_openai_api_key>
+OPENAI_BASE_URL=https://api.openai.com/v1  # Optional, can override
+OPENAI_ORGANIZATION=  # Optional
+
+# -----------------------------------------------------------------------------
+# OpenAI-Compatible Configuration (OpenRouter, Together.ai, Groq, etc.)
+# -----------------------------------------------------------------------------
+OPENAI_COMPATIBLE_BASE_URL=https://api.openrouter.ai/v1
+OPENAI_COMPATIBLE_API_KEY=<your_api_key>
+OPENAI_COMPATIBLE_MODEL=  # Optional, override CHAT_MODEL if needed
+
+# =============================================================================
+# External API Keys
+# =============================================================================
 ASKNEWS_CLIENT_ID = <your_asknews_client_id>
 ASKNEWS_CLIENT_SECRET = <your_asknews_client_secret>
 SERPAPI_API_KEY = <your_serpapi_api_key>
+
+# SearxNG Configuration
 SEARX_INSTANCE_URL = <your_searx_instance_url>
 SEARX_TOP_K_RESULTS = 5 # Number of top search results to retrieve

PROVIDER_EXAMPLES.md

@@ -0,0 +1,259 @@
+# LLM Provider Configuration Examples
+
+This guide shows you how to configure different LLM providers with PersonalAgent.
+
+## Supported Providers
+
+1. **Ollama** (local/remote)
+2. **LM Studio** (local)
+3. **OpenAI** (API)
+4. **OpenAI-Compatible** (OpenRouter, Together.ai, Groq, Fireworks, etc.)
+
+## Quick Start
+
+1. Copy `.env.template` to `.env`
+2. Set your `LLM_PROVIDER` and configure the relevant section
+3. Run the agent!
+
+---
+
+## Configuration Examples
+
+### 1. Ollama (Local)
+
+Default setup for local Ollama installation:
+
+```bash
+LLM_PROVIDER=ollama
+EMBEDDING_PROVIDER=ollama
+
+CHAT_MODEL=gpt-oss:120b-cloud
+EMBED_MODEL=nomic-embed-text
+TEMPERATURE=0.7
+
+OLLAMA_BASE_URL=http://localhost:11434
+OLLAMA_API_KEY=  # Leave empty for local
+```
+
+### 2. Ollama (Remote/API)
+
+Connect to a remote Ollama instance:
+
+```bash
+LLM_PROVIDER=ollama
+EMBEDDING_PROVIDER=ollama
+
+CHAT_MODEL=llama3.1:70b
+EMBED_MODEL=nomic-embed-text
+TEMPERATURE=0.7
+
+OLLAMA_BASE_URL=http://your-server.com:11434
+OLLAMA_API_KEY=your_api_key_if_needed
+```
+
+### 3. LM Studio (Local)
+
+Use LM Studio's local server:
+
+```bash
+LLM_PROVIDER=lmstudio
+EMBEDDING_PROVIDER=lmstudio
+
+CHAT_MODEL=llama-3.1-8b  # Or whatever model you have loaded
+EMBED_MODEL=nomic-embed-text
+TEMPERATURE=0.7
+
+LMSTUDIO_BASE_URL=http://localhost:1234/v1
+LMSTUDIO_API_KEY=  # LM Studio uses a dummy key
+```
+
+**Note:** Make sure LM Studio's local server is running before starting the agent.
+
+### 4. OpenAI
+
+Use OpenAI's official API:
+
+```bash
+LLM_PROVIDER=openai
+EMBEDDING_PROVIDER=openai
+
+CHAT_MODEL=gpt-4o
+EMBED_MODEL=text-embedding-3-small
+TEMPERATURE=0.7
+
+OPENAI_API_KEY=sk-your-openai-api-key
+OPENAI_BASE_URL=https://api.openai.com/v1  # Optional
+OPENAI_ORGANIZATION=  # Optional
+```
+
+### 5. OpenRouter
+
+Use OpenRouter to access multiple models:
+
+```bash
+LLM_PROVIDER=openai_compatible
+EMBEDDING_PROVIDER=openai_compatible
+
+CHAT_MODEL=anthropic/claude-3.5-sonnet
+EMBED_MODEL=text-embedding-3-small
+TEMPERATURE=0.7
+
+OPENAI_COMPATIBLE_BASE_URL=https://openrouter.ai/api/v1
+OPENAI_COMPATIBLE_API_KEY=sk-or-v1-your-api-key
+```
+
+**Popular OpenRouter models:**
+- `anthropic/claude-3.5-sonnet`
+- `meta-llama/llama-3.1-405b-instruct`
+- `google/gemini-pro-1.5`
+- `openai/gpt-4o`
+
+### 6. Together.ai
+
+Use Together.ai for fast inference:
+
+```bash
+LLM_PROVIDER=openai_compatible
+EMBEDDING_PROVIDER=openai_compatible
+
+CHAT_MODEL=meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo
+EMBED_MODEL=togethercomputer/m2-bert-80M-8k-retrieval
+TEMPERATURE=0.7
+
+OPENAI_COMPATIBLE_BASE_URL=https://api.together.xyz/v1
+OPENAI_COMPATIBLE_API_KEY=your_together_api_key
+```
+
+### 7. Groq
+
+Use Groq for ultra-fast inference:
+
+```bash
+LLM_PROVIDER=openai_compatible
+EMBEDDING_PROVIDER=openai  # Groq doesn't have embeddings, use OpenAI
+
+CHAT_MODEL=llama-3.1-70b-versatile
+EMBED_MODEL=text-embedding-3-small
+TEMPERATURE=0.7
+
+OPENAI_COMPATIBLE_BASE_URL=https://api.groq.com/openai/v1
+OPENAI_COMPATIBLE_API_KEY=gsk_your_groq_api_key
+
+# For embeddings
+OPENAI_API_KEY=sk-your-openai-api-key
+```
+
+**Popular Groq models:**
+- `llama-3.1-70b-versatile`
+- `llama-3.1-8b-instant`
+- `mixtral-8x7b-32768`
+- `gemma2-9b-it`
+
+### 8. Fireworks.ai
+
+Use Fireworks.ai for fast model hosting:
+
+```bash
+LLM_PROVIDER=openai_compatible
+EMBEDDING_PROVIDER=openai_compatible
+
+CHAT_MODEL=accounts/fireworks/models/llama-v3p1-70b-instruct
+EMBED_MODEL=nomic-ai/nomic-embed-text-v1.5
+TEMPERATURE=0.7
+
+OPENAI_COMPATIBLE_BASE_URL=https://api.fireworks.ai/inference/v1
+OPENAI_COMPATIBLE_API_KEY=fw_your_fireworks_api_key
+```
+
+### 9. Mixed Providers
+
+You can use different providers for chat and embeddings:
+
+```bash
+# Use Groq for fast chat, OpenAI for embeddings
+LLM_PROVIDER=openai_compatible
+EMBEDDING_PROVIDER=openai
+
+CHAT_MODEL=llama-3.1-70b-versatile
+EMBED_MODEL=text-embedding-3-small
+TEMPERATURE=0.7
+
+# Groq for chat
+OPENAI_COMPATIBLE_BASE_URL=https://api.groq.com/openai/v1
+OPENAI_COMPATIBLE_API_KEY=gsk_your_groq_api_key
+
+# OpenAI for embeddings
+OPENAI_API_KEY=sk-your-openai-api-key
+```
+
+---
+
+## Model Recommendations
+
+### Chat Models
+
+**For Quality:**
+- `gpt-4o` (OpenAI)
+- `claude-3.5-sonnet` (OpenRouter)
+- `llama-3.1-405b-instruct` (OpenRouter/Together.ai)
+
+**For Speed:**
+- `llama-3.1-8b-instant` (Groq)
+- `llama-3.1-70b-versatile` (Groq)
+- `mixtral-8x7b-32768` (Groq)
+
+**For Local:**
+- `llama3.1:70b` (Ollama)
+- `mistral:7b` (Ollama)
+- `qwen2.5:32b` (Ollama)
+
+### Embedding Models
+
+**OpenAI:**
+- `text-embedding-3-small` (fast, cheap)
+- `text-embedding-3-large` (best quality)
+
+**Local (Ollama):**
+- `nomic-embed-text` (recommended)
+- `mxbai-embed-large`
+- `all-minilm`
+
+---
+
+## Troubleshooting
+
+### "Connection refused" errors
+- Check that your local server (Ollama/LM Studio) is running
+- Verify the BASE_URL is correct
+- Check firewall settings
+
+### "Invalid API key" errors
+- Double-check your API key is correct
+- Make sure there are no extra spaces
+- Verify the API key has proper permissions
+
+### "Model not found" errors
+- Ensure the model name matches exactly (case-sensitive)
+- For Ollama: run `ollama list` to see available models
+- For LM Studio: check the model is loaded in the UI
+
+### Import errors
+- Run `pip install -r requirements.txt` to install all dependencies
+- Make sure `langchain-openai` is installed for OpenAI-compatible providers
+
+---
+
+## Cost Considerations
+
+**Free/Local:**
+- Ollama (local)
+- LM Studio (local)
+
+**Pay-as-you-go:**
+- OpenAI: $$$ (most expensive, best quality)
+- OpenRouter: $$ (varies by model)
+- Together.ai: $ (affordable, fast)
+- Groq: $ (very fast, affordable)
+- Fireworks.ai: $ (affordable)
+
+**Tip:** Start with Groq's free tier or use local Ollama for development!

agent.py

@@ -1,4 +1,3 @@
-from langchain_ollama import ChatOllama
 from langchain.agents import create_agent
 from langchain_core.runnables import RunnableConfig
 from tools.wiki import wiki_search
@@ -22,14 +21,16 @@
 import os
 from utils.spinner import Spinner
 from utils.markdown_render import render_markdown
+from utils.llm_config import get_chat_model
 from langgraph.errors import GraphRecursionError
 
 
 def main():
     x = input("Ask Kurama 🦊\n")
     load_dotenv()
 
-    llm = ChatOllama(model="gpt-oss:120b-cloud", temperature=0.7)
+    # Get chat model based on configured provider
+    llm = get_chat_model()
     recursion_limit = 100
     config = RunnableConfig(tags=["debug", "local"], recursion_limit=recursion_limit)
     agent = create_agent(

requirements.txt

@@ -2,14 +2,15 @@ langchain>=0.3.0
 langchain-core>=0.3.0
 langchain-community>=0.3.0
 langchain-ollama>=0.1.0
+langchain-openai>=0.2.0
 langgraph>=0.1.0
 
 duckduckgo-search>=6.3.0
 ddgs>=1.1.6
 wikipedia>=1.4.0
 requests>=2.31.0
 
-chromadb>=0.5.5        
+chromadb>=0.5.5
 python-dotenv>=1.0.1
 pydantic>=2.8.2
 tqdm>=4.66.1

retriever.py

@@ -1,10 +1,14 @@
-from langchain_ollama import OllamaEmbeddings
 from store.chromadb import collection
 from langchain.tools import tool
 import uuid
 from utils.spinner import Spinner
+from utils.llm_config import get_embedding_model
+from dotenv import load_dotenv
 
-embed = OllamaEmbeddings(model="nomic-embed-text")
+load_dotenv()
+
+# Get embedding model based on configured provider
+embed = get_embedding_model()
 
 @tool
 def add_to_db(text: str, metadata: dict = None):

tools/summarize_text.py

@@ -1,17 +1,21 @@
-from langchain_ollama import ChatOllama
 from langchain.tools import tool
 from utils.spinner import Spinner
+from utils.llm_config import get_chat_model
+from dotenv import load_dotenv
+
+load_dotenv()
 
 @tool
 def summarize_text(text: str) -> str:
     """
-    Summarizes a given text using the local Ollama LLM.
+    Summarizes a given text using the configured LLM provider.
     Useful for summarizing research reports, markdowns, or long notes.
     """
     s = Spinner("Running summarize_text…")
     s.start()
     try:
-        llm = ChatOllama(model="gpt-oss:120b-cloud", temperature=0.7)
+        # Get chat model based on configured provider
+        llm = get_chat_model()
         prompt = (
             "Summarize the following text in clear, structured Markdown format. "
             "Keep all essential details, and include a 'Key Takeaway' section at the end:\n\n"