Introduction
Woise is a platform for building voice agents that make and answer real phone calls. You assemble an agent on a canvas, connect a phone number to it, and it holds a conversation — listening, deciding what to say, and acting on what it hears.
Nothing about that requires a telephony engineer or a machine-learning team. The speech recognition, the language model, the speech synthesis and the turn-taking that decides when the agent should speak are all wired together for you. What you supply is the conversation: what the agent is for, what it should ask, and what it should do with the answers.
What runs on a call
Four things happen between a caller speaking and the agent replying, and it is worth knowing their names because every latency question and every quality question lands on one of them.
- Speech to text turns the caller's audio into words as they speak, not after they finish.
- A language model reads those words in the context of the current node's prompt and decides what to say and what to do.
- Text to speech turns the reply back into audio in the voice you chose.
- Turn-taking decides the moment the agent starts talking — whether a pause means "your turn" or "I'm still thinking". Get this wrong and an otherwise excellent agent interrupts people.
The first three are swappable per organization. The fourth is not something you configure so much as something you tune, through the interruption settings on individual nodes.
How an agent is put together
An agent is a workflow: a graph of nodes joined by edges. Each node is one step in a conversation — a thing to say, a question to ask, a decision to make, a system to call. Each edge is the rule for moving on: the model judging that a caller's goal is met, a keypad digit, or a condition on something the agent has already collected.
A run is one execution of that graph against one caller. It carries the transcript, the recording, the cost, and the outcome, and it is what every report and every log entry is about.
Nothing sits behind the graph. What you see on the canvas is what runs on the call.
A workflow decides what happens. A number gives it somewhere to be reached. And a region, fixed when you sign up, decides which telephony path is open to you, which compliance rules apply, and which currency you are billed in. Everything else — voices, tools, knowledge, integrations — changes what the agent can do during a call, not whether the call can happen.
Design and configure
Everything that shapes what the agent does while someone is on the line.
| Goal | Guide | What it covers |
|---|---|---|
| Learn the vocabulary | Concepts | Workflow, run, node, edge, region — the words the rest of these pages assume |
| Place a first call | First call | New account to a real transcript, in five steps |
| Assemble a conversation | Builder | Adding nodes, drawing edges, testing without a phone number |
| Choose the right step | Nodes | All ten node types, what each holds, and when to reach for it |
| Decide what it says | Prompts | Per-node prompts, the global prompt, and why one big prompt fails |
| Pick how it sounds | Voices | Voice selection and the trade-off against response time |
| Act during a call | Tools | Calling out to your systems while the caller waits |
| Ground its answers | Knowledge base | Documents the agent can consult instead of guessing |
| Change a live agent safely | Versions and forks | Editing safely while calls are live |
Connect and deploy
Getting an agent onto the phone network and into your other systems.
| Goal | Guide | What it covers |
|---|---|---|
| Attach a phone number | Numbers | The sourcing paths, which differ by region |
| Use your own carrier | BYO telephony | The seven supported carriers and the credentials each needs |
| Connect a PBX or SIP trunk | SIP | What a SIP trunk is, and both directions of call flow |
| Reach your other systems | Integrations | The catalog, and how authentication is handled for you |
| Answer incoming calls | Inbound | Routing an incoming number to a workflow |
| Call out to a list | Campaigns | Outbound batches, sources, and pacing |
Monitor and operate
What every call leaves behind, and how it is paid for.
| Goal | Guide | What it covers |
|---|---|---|
| Hear and read a call | Recordings | Where a finished run's audio and transcript live |
| See how agents perform | Reports | Outcomes across runs, built from what your agents extract |
| Track consumption | Usage | What is metered, and how to see it before the invoice |
| Understand the bill | Billing | How usage becomes an amount, and in which currency |
| Configure the organization | Settings | Org-level settings, including compliance configuration |
| Automate from your own code | API keys | Programmatic access and how keys are scoped |
Regions
Your region is fixed at signup. It is not a display preference — it decides which telephony path is open to you, which compliance rules you are subject to, and which currency you are billed in.
| Region | Guide | What it covers |
|---|---|---|
| All regions | Overview | What the choice affects, and what it cannot be changed to later |
| United States | United States | BYO-first number sourcing and the US compliance posture |
| India | India | Managed +91 numbers and the KYC requirement |
Start here
If you are new, read what Woise is for the shape of the product, then work through your first call — it takes an empty account to a real recorded conversation and takes about fifteen minutes.
Common questions
What is Woise?
Woise is a managed platform for voice AI agents. You design an agent in a drag-and-drop workflow editor, connect a phone number, and it handles real inbound and outbound calls over the phone network or in the browser. The infrastructure, carrier setup and scaling are run for you.
Do I need a phone number to start?
You bring a number you already own and import it — that is the primary path today, and it means you keep your carrier relationship and your existing number. Agents can also take and place calls in the browser over WebRTC with no number at all.
How does billing work?
You are billed for the call minutes your agents actually use, metered as credits against your account. There is no seat licence and no per-agent fee, so an agent that takes no calls costs nothing to keep.
Can an agent connect to my CRM or calendar?
Yes. Agents can read from and write to your CRM, calendar and support tools through the integrations marketplace — you authorise each connection once with OAuth, and no credentials are stored in your workflow.
Do I need engineers to build an agent?
The workflow editor is visual, so building and changing an agent needs no code. There is also a REST API, a Python SDK and a TypeScript SDK if you would rather drive the platform programmatically, and an MCP server so coding agents can build workflows for you.
What happens to call recordings and transcripts?
Recordings and transcripts belong to your workspace and are available through the dashboard and the API. Raw transcripts, recordings and phone numbers are never handed to third-party analytics or error tracking without redaction.
What infrastructure do I have to run?
None. Woise runs the speech models, the language model turn, the call recording pipeline and the telephony legs for you, and scales them with your call volume. There is nothing to provision, patch or keep running.
Still stuck? Talk to our team.