deepgram-go-speech-to-text

Using Deepgram Speech-to-Text from the Go SDK

When to use this product

Use this skill for pkg/client/listen work:

• prerecorded transcription with FromURL, FromFile, or FromStream
• live transcription with pkg/client/listen/v1/websocket
• channel-based or callback-based streaming flows

Use a different skill when:

• you need TTS output (deepgram-go-text-to-speech)
• you need text analysis on plain text (deepgram-go-text-intelligence)
• you need analytics overlays like summaries, topics, or sentiments (deepgram-go-audio-intelligence)
• you need Flux / conversational STT v2 (deepgram-go-conversational-stt)

Authentication

Set DEEPGRAM_API_KEY before constructing clients.

export DEEPGRAM_API_KEY="your_api_key"

This SDK reads env-backed defaults via the client option layer. Prefer API key or token auth supported by the repo's client options; do not hardcode credentials.

Quick start

Prerecorded REST:

package main

import (
	"context"
	"fmt"
	"log"

	api "github.com/deepgram/deepgram-go-sdk/v3/pkg/api/listen/v1/rest"
	listen "github.com/deepgram/deepgram-go-sdk/v3/pkg/client/listen"
	interfaces "github.com/deepgram/deepgram-go-sdk/v3/pkg/client/interfaces"
)

func main() {
	if err := run(); err != nil {
		log.Fatal(err)
	}
}

func run() error {
	ctx := context.Background()

	client := listen.NewRESTWithDefaults()
	dg := api.New(client)

	resp, err := dg.FromURL(
		ctx,
		"https://dpgr.am/spacewalk.wav",
		&interfaces.PreRecordedTranscriptionOptions{
			Model:       "nova-3",
			SmartFormat: true,
			Punctuate:   true,
		},
	)
	if err != nil {
		return err
	}

	fmt.Println(resp.Results.Channels[0].Alternatives[0].Transcript)
	return nil
}

Live WebSocket with channel fan-out:

package main

import (
	"context"
	"fmt"
	"log"

	listenws "github.com/deepgram/deepgram-go-sdk/v3/pkg/api/listen/v1/websocket"
	listen "github.com/deepgram/deepgram-go-sdk/v3/pkg/client/listen"
	interfaces "github.com/deepgram/deepgram-go-sdk/v3/pkg/client/interfaces"
)

func main() {
	if err := run(); err != nil {
		log.Fatal(err)
	}
}

func run() error {
	ctx := context.Background()
	handler := listenws.NewDefaultChanHandler()

	conn, err := listen.NewWSUsingChanWithDefaults(
		ctx,
		&interfaces.LiveTranscriptionOptions{Model: "nova-3", InterimResults: true},
		handler,
	)
	if err != nil {
		return err
	}
	defer conn.Stop()

	if ok := conn.Connect(); !ok {
		return fmt.Errorf("connect failed")
	}

	conn.Start()

	// The handler receives Open/Message/Metadata/UtteranceEnd events.
	// In a real app, stream PCM/audio chunks from your mic or file reader here.
	// For example (pseudo-code):
	//   for chunk := range audioChunks {
	//       if err := conn.WriteBinary(chunk); err != nil { return err }
	//   }
	//
	// When the input stream ends, flush any trailing audio and close cleanly:
	//   if err := conn.Finalize(); err != nil { return err }

	return nil
}

Key parameters

• interfaces.PreRecordedTranscriptionOptions
  • common fields: Model, Language, Punctuate, SmartFormat, Diarize, DiarizeModel (batch diarization version: latest/v1/v2), Redact, Utterances
  • use with pkg/api/listen/v1/rest: api.New(client).FromURL, FromFile, FromStream
• interfaces.LiveTranscriptionOptions
  • common fields: Model, Language, Encoding, SampleRate, Channels, InterimResults, Endpointing
• constructor families
  • REST: listen.NewRESTWithDefaults(), listen.NewREST(apiKey, options)
  • WS callbacks: listen.NewWSUsingCallback...
  • WS channels: listen.NewWSUsingChan...
• lifecycle
  • Connect() returns bool; call Start(), stream/write audio, KeepAlive() as needed, Finalize(), then defer conn.Stop()

API reference (layered)

1. In-repo reference
   • README.md
   • docs.go
   • pkg/client/listen/client.go
   • pkg/client/listen/v1/rest/client.go
   • pkg/client/listen/v1/websocket/client_callback.go
   • pkg/client/listen/v1/websocket/client_channel.go
   • pkg/client/interfaces/v1/types-prerecorded.go
   • pkg/client/interfaces/v1/types-stream.go
2. OpenAPI
   • https://developers.deepgram.com/openapi.yaml
3. AsyncAPI
   • https://developers.deepgram.com/asyncapi.yaml
4. Context7
   • /llmstxt/developers_deepgram_llms_txt
5. Product docs
   • https://developers.deepgram.com/reference/speech-to-text/listen-pre-recorded
   • https://developers.deepgram.com/reference/speech-to-text/listen-streaming
   • https://developers.deepgram.com/docs/speech-to-text

Gotchas

1. This repo uses listen package names for STT v1, not transcription.
2. Streaming code is split into callback and channel variants; copy the style that matches the surrounding package.
3. For WebSockets, pass a handler into NewWSUsingChan..., keep defer conn.Stop() near construction, and finalize before shutdown.
4. Live and prerecorded option structs are different; do not assume analytics-only prerecorded fields exist in live mode.
5. Use context.Context and return error; do not translate examples into exception-style control flow.

Example files in this repo

• examples/speech-to-text/rest/url/main.go
• examples/speech-to-text/rest/file/main.go
• examples/speech-to-text/websocket/microphone_channel/main.go
• examples/speech-to-text/websocket/microphone_callback/main.go
• tests/edge_cases/keepalive/main.go
• tests/edge_cases/reconnect_client/main.go

Central product skills

For cross-language Deepgram product knowledge — the consolidated API reference, documentation finder, focused runnable recipes, third-party integration examples, and MCP setup — install the central skills:

npx skills add deepgram/skills

This SDK ships language-idiomatic code skills; deepgram/skills ships cross-language product knowledge (see api, docs, recipes, examples, starters, setup-mcp).