Create a Conversational Voice AI Agent in the Speechmatics Portal

Learn how to build a real time voice AI agent using Speechmatics Portal. From creating an agent in the Portal to adding system instructions, Custom Dictionary, and Function Calling (we wire up a “create reminder” tool), then test it live in the browser. Sign up to the Portal: https://portal.speechmatics.com/signup/ Flow overview (Voice Agents): https://docs.speechmatics.com/voice-agents-flow Flow WebSocket API reference: https://docs.speechmatics.com/api-ref/flow-voice-ai-websocket Function Calling in Flow: https://docs.speechmatics.com/voice-agents-flow/features/function-calling Flow setup (template_id, variables): https://docs.speechmatics.com/voice-agents-flow/setup React Native guide (full code): https://docs.speechmatics.com/voice-agents-flow/guides/react-native Python client (Flow): https://github.com/speechmatics/speechmatics-flow 00:00 Intro — What we’re building (Flow agent) 00:33 Create a new agent (starter template) 01:20 Choose language & voice (incl. Arabic / British options) 02:04 Add system instructions / persona 02:36 Add items to Custom Dictionary 03:00 Set the first message (welcome prompt) 03:24 Add Tools via Function Calling (reminder schema) 03:46 Save the agent & test in Portal 05:28 Live conversation with the agent 07:48 Start a new session 08:31 Tool invocation demo (set a reminder) 09:13 Next steps & wrap-up