What makes CapeChat different from other products powered by the ChatGPT API?
CapeChat is uncompromisingly focused on data privacy and security!
Without CapeChat, the data you send to OpenAI and other intermediaries is at risk of being observed, stored, or inadvertently leaked in plaintext. CapeChat ensures that your data remains protected at all times using encryption and advanced privacy and security techniques.
CapeChat redacts sensitive personal, financial, and health information from your documents and messages before they are processed by the OpenAI ChatGPT API.
What file types do you support for documents?
CapeChat supports many popular filetypes for documents, including:
- PDF (.pdf)
- Word (.docx)
- Excel (.xslx)
- PowerPoint (.pptx)
- CSV (.csv)
- Text (.txt)
- Markdown (.md)
Can I chat with multiple documents?
Yes, you can upload multiple documents to a single chat. You can also start a new chat and upload a new set of documents, keeping your conversations focused on a particular topic.
Who can see my documents and messages?
Only you can see your documents and messages.
- Cape keeps your data protected at all times.
- Your sensitive data is only ever visible to you.
- You can delete your data at any time.
Remarkably, your data is even protected from the humans at Cape Privacy and our sub-processors at AWS and OpenAI:
- Cape Privacy and AWS can never see your data.
- OpenAI only processes your redacted data.
How does CapeChat ensure the highest level of data security and privacy when dealing with sensitive documents?
- Servers are protected with Intel® Total Memory Encryption and AWS Nitro Enclaves.
- Storage is encrypted with AES 256 and RSA 4096.
- Encryption keys are protected by AWS KMS and only accessible to AWS Nitro Enclaves.
- Sensitive data is protected with state-of-the-art de-identification of PII, PCI, and PHI before being processed by OpenAI’s ChatGPT API.
- Network is encrypted with TLS.
Do you have an API to integrate CapeChat?
Yes. You can upgrade to CapeChat Plus to get access to the API.
What are some real-life use cases where CapeChat has shown a positive impact in handling sensitive data?
At Cape, we are thrilled to be bringing the transformative capabilities of GPT-4 to companies that wish to be good stewards of the sensitive data they are entrusted with.
- Financial Services: CapeChat streamlines the interpretation of financial data from disparate sources and calculates financial formulas to reduce time-to-decision.
- Security & Compliance: CapeChat performs automated question answering derived from sensitive policy documents, improving trust and saving manual effort.
- Accounting: CapeChat automates the conversion of proprietary banking statements into structured bookkeeping data.
How easy is it to integrate CapeChat with existing systems and tools used for document storage and collaboration (e.g., Google Docs, GitHub, Slack, and Zoom)?
Our API makes it easy to integrate into your existing tools and workflows. You can use the API to build custom integrations with your own applications, databases, file storage, and productivity tools. You must upgrade to CapeChat Plus to access the API.
Can CapeChat also be used for non-sensitive data?
While CapeChat can be used with any data, it is most applicable to use cases where privacy is of concern.
What are the restrictions on the free version of CapeChat?
The free version of CapeChat is intended to be a way to get familiar with the functionality provided by the Cape API. You can upload documents to create a secure custom knowledge base, then interact with it by asking questions. CapeChat will respond with answers using the power of the ChatGPT API. Your data is always treated as confidential, stored in encrypted format, and cleansed before being processed by OpenAI. With CapeChat Plus, you can enable GPT-4, have full control over prompts, and direct access to the Cape API to build your own integrations.
Can you explain the per-token, usage-based pricing model for the Cape API, and how it benefits users compared to traditional pricing models?
Tokens are sequences of characters that represent a unit of meaning to the ChatGPT API. While it depends on the specific text, 100 tokens can be roughly estimated as 75 english words. Pricing in this fashion means that cost is directly proportional to level of usage (e.g. amount of text), rather than a flat rate or a set number of requests. This is intended to be both cost effective and fair for all levels of usage.