Upgrade Your ChatGPT-4 Conversations with Semantic NetFlow
Unlock the full potential of your private documents and research collections.
Zero ChatGPT-4 API fees. Painless setup. Maximum relevance. $20/month, period.
Introducing Semantic NetFlow from cogniant.ai, a groundbreaking solution for interacting with web-based Large Language Models (LLMs) like ChatGPT-4/3.5. Our easy-to-use solution supercharges your LLM chat experiences by intelligently analyzing the content and sentiment of the prompts that you send to ChatGPT, then transparently rewriting each intercepted prompt with relevant context memory extracted from document embedding Collections that you create.
Seamless Integration with the ChatGPT-4/3.5 Web UI
Semantic NetFlow works as a simple web browser plugin which intercepts and modifies web UI-based LLM chat prompts by analyzing the content of your chats with ChatGPT-4/3.5 on-the-fly, then enriching each conversational turn with relevant and targeted context discovered from your local document Collections. Just pull a Docker container on WSL 2 for Windows, Linux, or MacOS, and then import your document Collections by dragging and dropping curated documents such as PDFs, Word documents, text files, JSON dumps, and other supported document file types directly into your locally-deployed Semantic NetFlow proxy container. We are releasing a hosted solution as well, which requires only the browser plugin and a cogniant.ai API key to access your document Collections stored securely on Amazon's AWS cloud.
High Performance, Locally-hosted Vector Similarity Search
Our platform is based around a high-performance, semantic textual similarity vector database that is used for indexing, clustering, and semantic searching across your locally-stored document embeddings. Supported document types such as [ .pdf | .docx | .csv | .txt | .json ] are simply dragged and dropped into the Semantic NetFlow container web UI for import and storage, at which point those document contents are indexed, searched, retrieved, and then encapsulated as context memory within any future ChatGPT completions that are semantically similar to the content of your chats.
No More Expensive API Subscription Fees
Semantic NetFlow is compatible with the free version of ChatGPT, and requires no OpenAPI embeddings API support or associated fees. Our solution is 100% self-hosted and with a hosted version in the works, with all network traffic analyzed and then modified locally, prior to submission to remote LLM endpoints. All you need is a browser plugin, then drag-and-drop your documents for the embeddings and vector database import process, making your document Collection(s) accessible by ChatGPT as context memory during each conversational turn. $20/month, no hidden fees.
Enhance your ChatGPT-4 experience with access to your hand-annotated document Collections, enabling you to generate marketing research, detailed scientific citations, precise legal pleadings, big data question answering and summarization, exact and precise generative code generation including code refactoring and function/class documentation, and much more, all with the power of ChatGPT-4/3.5 web UI in tandem with Semantic NetFlow. Each Collection you create is simply a folder that you drag-and-drop documents into, all of which becomes a part of a high dimensional cluster of features that are injected as memory artifacts into your web-based LLM conversations.
Just install our browser plugin on [ Chrome | Microsoft Edge | Safari | Opera Next | Firefox ], and start enjoying an upgraded ChatGPT-4 experience, without any OpenAI plumbing or exorbitant API fee requirements. Share the benefits of your document embedding Collections with all users on your local area network by allowing them to chat with OpenAI using your imported document Collections, and soon, allow others to chat with your document Collections remotely via secure FQDN Tailscale Funnel web proxy and API endpoints, even if your containers have been deployed behind firewalls or NAT devices, and without any Tailscale client requirements for securely accessing those remote resources.
- $20 per month, flat rate and with no hidden fees. Pay-as-you-go hosted version being released, based on AWS
- Network calls to OpenAI LLM resources are analyzed and then augmented/rewritten with vector database search contents prior to transmission
- Collection-based prompt engineering for precise control over ChatGPT interactions, each of your imported Collections can be customized with specific LLM instructions for how the data should be interpreted
- Automated document retrieval and LLM network traffic modification without user intervention, everything happens locally, transparently and according to user-configurable rules on how vector similarity search data are queried
- Secure on-site storage of documents without costly API fees. You own, manage, and are the custodian of your own documents and resulting Collections. No document embedding vendor lock-in with APIs that have expiration dates
- Compatible with web-based LLMs such as ChatGPT-4/3.5-turbo that are accessible on the OpenAI free service tiers
- Ensembling methods in the works, to utilize semantic textual similarity model pipelines with intercepted ChatGPT prompts, to determine which Collection(s) should be queried according to prompt context
- Experimental browser automation support via Selenium, to expose a similar set of programmatic endpoints on the local Docker instance as what are offered by OpenAI and Azure paid API services
Infinite Possibilities with Your Curated Collections
Easy Setup and Secure Sharing
Key Features and Functionality
