Connected sources
Connects to any AI source you approve.
Any AI source you approve, including providers, chat assistants, applications and agents. The sources connected in your deployment are agreed during scoping.
Source types
The kinds of sources PromptLeash reads.
AI assistants and chat
General-purpose assistants your people use for everyday work.
For example
Microsoft 365 Copilot, ChatGPT Enterprise, Claude, Gemini
Model providers and APIs
The model platforms your teams and applications call directly.
For example
OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock, Azure OpenAI
Coding assistants
AI tools that help engineers write, review and change code.
For example
GitHub Copilot, Cursor, Claude Code
Agents and automation
Agents and automated workflows that use AI to act on your behalf.
For example
Microsoft Copilot Studio, agents built on agent frameworks
AI gateways and observability
Gateways and logging layers that already capture AI traffic across tools.
For example
LiteLLM, Portkey, Helicone
Your own AI applications
Applications your organisation has built on model APIs, from internal tools to customer-facing services.
Scope
You decide what is in scope.
- What does PromptLeash read?
- Approved AI activity from the sources you put in scope, including interaction content such as prompts, responses and agent activity, within the agreed scope. Nothing is analysed until it is in the agreed scope.
- Where is it processed?
- In your environment. Classification and aggregation run there, alongside your AI stack rather than in the path of a live request.
Product names are trademarks of their owners and are listed as examples of source types. Listing does not imply a partnership or endorsement.
Check your sources in a conversation.
Book a demo to discuss your priorities and see the current product. We then agree which data sources, roles and workflows are relevant, complete your security review, and deploy inside your environment.