Your First Agent in 5 Minutes
Build and talk to a working voice agent in five minutes using Web Calls — no telephony setup required.
This is the fastest path to a working voice agent. You'll create an agent, build a small conversation flow, and talk to it in your browser — no phone number, no telephony provider, no configuration.
Go to talkr.intelstacks.com/workflow. Self-hosting? Use your own instance's URL instead, and follow Enterprise Deployment first to get the platform running.
Step 1: Create an agent
On the Agents page, click Create Agent. You have two options:
Use Agent Builder
Describe your use case and Talkr generates a complete starting workflow for you — nodes, prompts, and pathways — ready to test immediately. The fastest way to see something working. Limited to 10 generations per organization per day.
Blank Canvas
Start from a single pre-filled Start Call node and build the graph yourself. Always available, no limit — use this if you'd rather build by hand or have hit the daily generation limit.
Either way, you land in the Voice Agent Builder with something on the canvas to work with.
Step 2: Check (or build) the flow
If you used Use Agent Builder, skim what was generated — a Start Call node, a couple of Agent nodes covering the stages of your use case, and an End Call node, connected by pathways. Read through the prompts; tweak anything that doesn't quite match what you had in mind.
If you used Blank Canvas, build a minimal flow yourself — three nodes are enough to get started, and you don't need the full graph model yet:
- Start Call — already on the canvas. Edit its prompt to greet the caller and describe what the agent should do, e.g. "Greet the caller, ask what they need help with, and try to book an appointment. Confirm the date and time before ending the call."
- Agent node — add one from the node panel and connect it from Start Call. This is where the main conversation logic lives.
- End Call node — add one and connect it from the Agent node so every path through the flow has somewhere to land.
The more specific your prompts, the better the agent performs — describe the goal, the questions to ask, and how to close the call.
Step 3: Talk to it with a Web Call
Click Try Your First Web Call. This runs the full pipeline — speech-to-text, LLM, text-to-speech — straight from your browser microphone, exactly like a real phone call would, just without a phone number.
Nothing to configure first: new organizations get free trial credits, so this call runs on Talkr's own LLM/STT/TTS keys automatically for any service you haven't configured yet. See Billing for how the balance works, or Model Configuration to bring your own keys instead.
While the call is active, you can watch:
- The live transcript as the conversation happens
- Node transitions as the agent moves through the graph
- Any tool calls the agent makes
Talk to it like a real caller would. Try to get it off-script — say something unexpected — and see how it responds.
Step 4: Iterate
Didn't go the way you wanted? Open the node whose behavior you want to change and edit its prompt directly. Save, then start a new Web Call to hear the change immediately. No redeploy, no restart.
Common first tweaks:
- Adjust tone or wording in an Agent node's prompt
- Add shared instructions (tone, objection handling) to the Global node
- Add a Webhook node to send call results somewhere once you're happy with the flow
Next Steps
You have a working agent that runs entirely in the browser. From here:
- Take real calls — connect a telephony provider and trigger calls via API Trigger or inbound routing.
- Learn the graph model in depth — see Voice Agent Builder for all node types and how pathways work.
- Give it tools — let the agent call external APIs or transfer calls with Tools.
- Embed it on a website — skip telephony entirely and let website visitors talk to the agent via Add to Website.