About GitHub
LLMStack is an open-source, no-code platform for building generative AI agents, workflows, and chatbots. It allows users to chain multiple LLMs together, integrate their own data from various sources (CSV, TXT, PDF, DOCX, PPTX, Google Drive, Notion, websites, direct uploads), and deploy either to the cloud (via Promptly) or on their own infrastructure. The platform includes a no-code builder, tools for connecting agents to external services, and support for triggering AI chains from Slack or Discord. It handles data preprocessing and vectorization automatically and provides an out-of-the-box vector database.
Key Features
Build generative AI agents (e.g., AI SDRs, Research Analysts, RPA Automations) without writing code
Chain multiple LLMs together to create complex generative AI applications
Use generative AI on your data from multiple sources (CSV, TXT, PDF, DOCX, PPTX, Gdrive, Notion, websites, direct uploads) with automatic preprocessing and vectorization
No-code builder for creating AI chains
Deploy to cloud (Promptly) or on your own infrastructure (self-hosted)
Trigger AI chains from Slack or Discord
Connects agents to internal/external tools, web search, and internet browsing
Pros & Cons
Pros
- No-code platform makes AI accessible without programming skills
- Open-source (MIT license) allows full control and self-hosting
- Supports multiple data sources and automatic preprocessing/vectorization
- Can chain multiple LLMs for more sophisticated applications
- Provides agents with tool-use capabilities (web search, internet browsing)
- Offers both cloud and on-premise deployment options
Cons
- Requires Docker for background job execution if self-hosting
- Default keys for LLM providers must be manually configured unless using cloud version
- Primarily designed for users with some technical comfort (pip install, Docker setup for self-hosted)
Best For
Building custom AI agents for sales, research, automationCreating complex AI workflows by chaining multiple language modelsDeveloping chatbots that use your own business dataAutomating repetitive tasks (RPA) with AI agentsIntegrating AI capabilities into Slack or Discord workflows
FAQ
What is LLMStack?
LLMStack is an open-source, no-code platform for building generative AI agents, workflows, and chatbots. It chains multiple LLMs and integrates with your own data.
How do I install LLMStack?
Install via pip: 'pip install llmstack'. Requires Docker for background jobs. Start with 'llmstack' and access at localhost:3000.
What data sources are supported?
LLMStack supports CSV, TXT, PDF, DOCX, PPTX, Google Drive, Notion, websites, and direct file uploads.
Can I deploy LLMStack on my own servers?
Yes, you can self-host LLMStack on your own infrastructure. A cloud version is also available via Promptly.