;
The following sections contain detailed technical information about the Planview Anvi, such as the models used and the architecture behind Anvi.
Planview Anvi uses Anthropic models via Amazon Bedrock APIs (AWS is the subprocessor). Anvi keeps up with the latest Anthropic model releases but on a delay of a couple of months; this helps ensure enough time for testing the latest models and for reserving quota availability.
Anvi uses the following models:
While these models are multimodal, Anvi only uses the text mode.
Anvi uses the same underlying models as Claude, as Claude is a web app, not a model. Therefore, Anvi shares the same underlying intelligence as Claude but differs in the application integration, data accessibility, and how Anvi is directed via prompting.
The following image shows the architecture on which Planview Anvi is built:
Planview Anvi Chat architecture
Ensure that the applicable domains in the following list are in the allow list for your organization:
Use HTTPS (port 443) for API endpoints.
| # |
Title |
Reasoning & Background | Impact Level if not Enabled |
|---|---|---|---|
| 1 |
Allow Anvi’s domain
The domain is based on your Planview account’s region. |
Used to communicate with the system |
Unable to use Anvi |
| 2 |
Allow cookies to be set on the browser |
Anvi relies on the AWSALBCORS cookie to route messages within the same conversation to the right place.
|
Severely impaired functionality - Anvi may forget information or be unable to register a user’s subsequent instructions. |
| 3 |
Allow LaunchDarkly |
Used for feature management |
Severely impaired functionality - users cannot use some features or fetch certain data. |
| 4 |
Disable buffering of Server-Sent-Events (SSE) for Anvi's domains above. |
Anvi uses SSE to send progress updates and receive action confirmations. |
Impaired functionality - attempts to use actions will timeout, and users will experience long waits without updates. |