FAQ

1. What security protocols does Codespell adhere to?

Codespell has been developed with robust security measures at its core. It employs  industry-standard TLS 1.2+ encryption for data in transit and AES-256 encryption for data at  rest. We also implement single sign on (SSO), and support role-based access control (RBAC).  In addition, Codespell is designed with strong data privacy guarantees and solid guardrails for  generative AI.

2. Which Large Language Models (LLMs) does Codespell use, and how are they  secured?

Codespell leverages state-of-the-art LLMs that offer extremely strict data privacy guarantees.  We use Azure OpenAI’s gpt4o for the Coding Assistant and AWS Bedrock’s Claude 3.5 for the  Design Studio. None of the data you trust Codespell with and none of the code Codespell  generates is used to train the LLM’s. If you are curious about the data guarantees from the  LLMs we use, please visit :
For Azure OpenAI - Data, privacy, and security for Azure OpenAI Service - Azure AI services | Microsoft Learn
For AWS Bedrock Claude - Secure Gen AI Apps - Amazon Bedrock Security and Privacy - AWS

3. Does Codespell store any customer data? If so, how is it encrypted?

Codespell does store the entity relationship diagrams, API testing documentation and IaC  setup you create as part of the Design Studio feature we offer. All of this data is encrypted by  AES-256 encryption at rest.

4. Does Codespell comply with industry standards?Do we have regulatory compliance  certifications?

Codespell complies with industry standard encryption standards and we are working towards  getting SOC 2 compliant. Every feature that we develop has stringent vendor penetration testing  performed before it is shipped.

5. Can I bring my own LLM to Codespell?

Yes, this feature is upcoming with a release date of March. For the AI coding assistant  feature, the plan will be to allow users to bring any model they would like. However, for the  Design Studio feature Codespell offers, we only allow users to bring the keys for the LLMs  Codespell uses (AWS Bedrock Claude 3.5). The reason for this restriction for Design Studio is  because of the rigorous testing that has gone to determine the best LLM that can help us meet  our SLA of offering consistent and standardized code every time.

6. What is Codespell’s unique differentiator compared to other coding assistants?

Codespell is not just a Coding Assistant. While we do support that feature, our primary  differentiator is our Design Studio that we believe will allow engineers to eventually automate  the entire Software Development LIfe Cycle. We already support grounds-up project building by  helping automate CRUD API generation, API Test generation and Infrastructure as Code from  design. Our ultimate goal is to help the software industry move from the current AI-assisted  landscape of code generation to an AI-driven landscape.

7. What languages does Codespell currently support? What IDE’s do we support?

Codespell’s AI Coding Assistant supports code generation in any language. Our design studio  offers these languages as of now:

• - CRUD API Generation: Java, Node.js, .Net, PHP, Golang, Python (upcoming)
• - API Test Script Generation: REST Assured based custom framework, Postman (upcoming)
• - Infrastructure as Code: Terraform (AWS), Google/Azure (upcoming)
• - Frontend Code Generation: FIGMA to Angular/React (upcoming)

Codespell integrates with Visual Studio Code, Visual Studio, IntelliJ and Eclipse IDEs.