"COMPLETED", "PENDING", "CANCELLED", and "ERROR".
"call").
"user" or "assistant").
"call" for now.
5. The total generations per conversation will be n\_permutations + 1 (one for the original user messages plus one per permutation).
GET /v1/node\_tests/run/:id to retrieve the full results.
Enable watchtower under the Compliance & Policy Management section page of the dashboard
This node uses watchtower to enable guard rails specifically on the loop condition, which will transfer the call if the condition fails after unreliable results are detected
Along with converting our node views into draggable panels, you can pop them out to create floating windows
At any point during a test chat, edit your user message to branch to a new conversation, where that input was used and subsequent agent responses are changed. You can view and swith between these from the conversation UI
View detailed logs including webhooks, knowledge base calls, and node transitions (routes and loops)
Reference previous calls using your pathway, and generate user responses based on a node's prompting
Run your prompting against the generated user responses and see how often your outputs meet your LLM grading goal
Describe changes in plain language directly from the pathway editor
Review every proposed change before applying to your pathway
Select groups of nodes within your pathway to be saved as a block
After selecting, click the save button in the bottom left to create a new block
Insert any block directly onto the canvas with one click through the "Add New Node" pane
Open Standards at individual nodes
Select a source call to automatically generate a simulation prompt and success definition
View pass and fail outcomes across all 10 test iterations
Define output fields and generate a transformation script from a real call
Run backtests against real calls in the editor
Open any node directly in the testbed from an active chat session
Edit the node prompt and configure simulation scenarios and success criteria side by side
Review simulation results and inspect individual conversation transcripts
Validate variable extraction with exact match evaluation across 10 simulations
Test loop conditions with a True/False success definition and configurable threshold
Start by choosing inbound or outbound — each is configured separately
Select your provider and enter your SIP server address
Manage all your trunks and view per-trunk call logs from the SIP dashboard
Port existing numbers with a guided LOA form — carrier details, authorized rep, and target port date
Pick a voice, type your text, and generate speech directly from the dashboard
Quick-start code examples and full API reference including streaming
What used to require a separate node for every tool call can now live inside a single Default node. Reference the name of your tool within the dialogue prompt to describe when it shoudld be triggered
Each tool is configured with its own name, webhook, and settings directly inside the node
Each tool is configured with its own name, webhook, and settings directly inside the node
You can also choose how the agent handles the request response. Decide to respond naturally and continue, or respond and follow defined pathways to route the call
Set a custom conversation directly on the standard instead of selecting one from your call history
Build out conversation history exactly as you see fit. With additional section for adding the conversation happening at that node (for variable extraction and loop condition standards)
Test the standard on simulation prompt and success definition
The Book Service Appointment node uses a Custom Tool to call the Cal.com Create Booking integration directly from the node
The Check Warranty node runs inline JavaScript to calculate whether the warranty is active, how many days remain, and the expiry date
The Create Support Ticket node uses a Tool Chain. Input variables collected earlier in the conversation are passed into the chain
Step 1 is a Code step that generates a ticket number and formats the full ticket payload, setting priority based on warranty status
Step 2 is a Webhook step that POSTs the formatted ticket and extracts the confirmed ticket number from the response
The full pipeline builder with input variables, code editor, and Norm open side by side
Configure a test scenario with a caller persona, call start prompt, and evaluation criteria to define what a successful outcome looks like
Review the full transcript of the agent interacting with the test caller, including the evaluation result
Toggle authentication on, and decide on how strict your agent acts to authenticate the caller (flexible vs sequential)
Configure methods like SMS code, which automatically send users an OTP over text messsage, or write custom validation logic with custom JS code
Decide when persona authentication takes place, and gate specific tools and pathways behind different verification methods
Include optional prompting guides for how your persona authentication handles opening, succuess, and failure scenarios
Configure what happens when verification fails: retry limits, cooldown periods, and whether to end the call, transfer, or continue
Multiple editors on the same pathway with live cursors and collaborator avatars visible on nodes and edges
When a collaborator makes changes, a banner lets you view the diff or refresh to sync
An interactive map of every source and question in your Knowledge Base, color-coded by coverage and resolution
*** ### Custom Skills for Norm Teach Norm to do exactly what your team needs. Custom skills are scoped to your organization and live alongside the built-in skill set in the Norm chat. * Type `/` in Norm chat to trigger any skill or to create a new one via "Create custom skill", with the editor opening in a side drawer with five fields: command, name, description, when to use, and the content Norm should follow * Use skills for anything your team does repeatedly, like enforcing a brand voice across dialogue prompts, scaffolding common flows such as insurance verification, or generating your standard QA scenarios for every new pathwayType `/` in Norm to open the slash menu and create or trigger a skill
Build the skill in a side drawer with command, name, description, when-to-use, and content fields
Status: {state}
Branded contact
Tapbacks and effects
Voice memos
Files and reports