What is Graphlit Not?

We are not just another OpenAI wrapper.

We can leverage OpenAI models, but we are agnostic to Large Language Models such that we can incorporate any API-exposed LLM within your data workflows.

Given our years of experience building media archives for film studios, video transcoding platforms for broadcasters, and data ingestion workflows for metadata and content, we bring a deep background of building scalable platforms which also prioritize API integrations with your applications and services.

We are not a set of Python packages to integrate.

If you are familiar with open source projects like LangChain or LlamaIndex, those libraries are a great 'bag of tools' when building your own Knowledge Applications.

But what they aren't is a cloud-native service, which provides secure content storage, recurrent data feeds, automated data workflows, data modelling, entity resolution, multi-tenant support and usage-based tracking and billing.

Graphlit is API-first - we manage a serverless cloud-native data platform for your applications to build on.

We are not another 'Chat with your PDF' application.

Graphlit is for developers. You can build any vertical AI application on top of Graphlit, using any multimedia data - from PDFs to audio recordings to videos or even CAD drawings. We aim to support the widest variety of data sources and formats.

We provide content management for a wide variety of content and file types, from Slack messages to large TIFF images to complex PDFs. We can automatically create renditions of your content, such as image thumbnails, streaming video previews, and convert file formats for use in your applications.

Graphlit supplies richer context for LLM prompts than just text from a document or audio transcript. We provide contextual knowledge based on the entities and relationships extracted from your content, in addition to technical metadata from the files, which lets us tune the prompt context for higher quality retrieval and conversational context.