AnythingLLM is really slow on my system. I am looking for a tool that can maintain context upto 40 chats. can refer to attached document on each chat. work with LM Studio server locally on Desktop. all this is required to create Short Novel for my self.

If you're looking for a tool that can maintain context across multiple chats (up to 40) while allowing for document attachments and local operation, you might consider the following options: ### 1. **Local AI Frameworks** - **LangChain**: This framework is designed for building applications with language models. It can manage context and can be integrated with local models or servers. You can configure it to maintain context between multiple interactions. It supports document handling as well. - **Haystack**: An open-source framework for NLP that offers support for document retrieval and can be configured to maintain conversation contexts. You can also use this framework locally. ### 2. **Local LLM Tools** - **Oobabooga's Text Generation Web UI**: If you are running your own LLM server (like LM Studio), you can use interfaces like Oobabooga's that allow for context management and can attach documents to each conversation. - **GPT-4 All**: A fine-tuned version of AI models that can run locally. It supports extending the model's capabilities to hold context across interactions and can handle multiple document references. ### 3. **Document Management and AI Integration** - You may want to integrate your writing workflow by using a note-taking or document management tool like **Obsidian** or **Notion** combined with an LLM. While these platforms don’t natively integrate with LLMs, you can use API calls to send and pull the context as needed. ### 4. **Development Environment** - You might also consider setting up a simple web application using **Flask** or **Streamlit**, which could act as an interface. You can integrate your local LLM and set it up to handle chat sessions and document attachments efficiently. ### Tips for Better Performance - **Optimizing Local Resources**: Ensure that your machine has enough resources (RAM, CPU/GPU) allocated for running LLMs efficiently. - **Batch Processing**: Reduce response times by batching inputs or processing multiple queries together. - **Configuration**: Make sure to tweak the parameters of your language model and server configuration for optimal performance based on your hardware. ### Conclusion While you explore these options, ensure that you check out their documentation for setup guidelines and API references. Creating a short novel using LLM while managing documents and context over a series of chats can significantly enhance your writing process. Consider your comfort with programming, as some of the solutions may require coding to implement fully.