recipes

# Deepgram Recipes Agent-maintained micro-recipes showing how to use every Deepgram SDK feature across every supported language. Each recipe is a focused, runnable snippet — not a full app. ## When to use recipes - You know the product and feature; you want the shortest working code - You want `example.py` / `example.js` / `example.go` / etc. you can copy into your project - You want a language-specific answer to "how do I call `{feature}` with the Deepgram SDK?" **Use a different skill when:** - You want a full starter app with a web UI, deploy config, etc. → `starters` skill - You want integration with a third-party platform (Twilio, LiveKit, Vercel AI SDK, Discord, etc.) → `examples` skill - You want the full API contract (params, responses, message shapes) → `api` skill - You want a shell command rather than application code → `cli` skill ## Browse recipes Repository: <https://github.com/deepgram/recipes> Coverage matrix: <https://github.com/deepgram/recipes/blob/main/COVERAGE.md> ## Recipe structure ``` recipes/{language}/{product}/{version}/{recipe}/ example.{ext} # runnable, < 50 lines, reads DEEPGRAM_API_KEY from env example_test.{ext} # runs the example as a subprocess, asserts output README.md # feature explanation, params, sample output, how to run ``` ## Products covered | Product | Recipe examples | |---|---| | Speech-to-Text: Nova (`/v1/listen`) | 26 recipes: transcribe-url, transcribe-file, streaming, streaming-file, punctuate, smart-format, paragraphs, utterances, diarize, multichannel, numerals, measurements, dictation, filler-words, profanity-filter, redact, replace, search, keywords, keyterm, detect-language, detect-entities, summarize, sentiment, topics, intents | | Speech-to-Text: Flux STT (`/v2/listen`) | Two recipes. `streaming` opens the `/v2/listen` WebSocket with `model=flux-general-en`, `encoding=linear16`, `sample_rate=16000` and prints `TurnInfo` events (transcript, turn index, event type) in place of v1 interim/final pairs. `transcribe-url` sets `model=flux-general-en` on the SDK's prerecorded transcribe-URL call with `smart_format`. The CLI has only `transcribe-url`. No recipe is dedicated to EOT or eager EOT thresholds, mid-session `Configure`, or keyterms | | Text-to-Speech: Aura (`/v1/speak`) | generate-audio, stream-audio, websocket-streaming, select-model, select-encoding, bit-rate | | Audio Intelligence (`/v1/listen`) | summarize, sentiment, topics, intents, entities | | Voice Agents | connect, custom-llm, custom-tts, function-calling | | Text Analysis (`/v1/read`) | summarize, sentiment, topics, intents | Audio Intelligence and Text Analysis run the same analysis on different inputs, so pick the row by what you already have. The Audio Intelligence recipes are query parameters layered on `/v1/listen`, so the input is audio and `entities` is one of them. The Text Analysis recipes are one `POST /v1/read`, so the input is text and there is no entities recipe: `detect_entities` on `/v1/read` returns 400 `{"err_code":"INVALID_QUERY_PARAMETER","err_msg":"unknown query parameter: detect_entities"}`. Open the `audio-intelligence` skill for the first and `text-intelligence` for the second. Text Analysis recipes exist in all seven languages. Nova is the general-purpose STT family; Flux STT is designed for conversational audio and voice agents. Both are actively maintained — see the `api` skill's "Nova vs Flux STT" section for the decision guide. Note that Flux STT (`/v2/listen`) and Flux TTS (`/v2/speak`) are separate products that share the Flux name. **Flux TTS (`/v2/speak`) has no recipes yet.** The TTS recipes above cover Aura on `/v1/speak` only — there is no `text-to-speech/v2` directory in the recipes repo. For Flux TTS, use the `api` skill's "Aura vs Flux TTS" section for the endpoint contract, or the `starters` skill's `flux-tts` apps (node, flask, fastapi, django, java) for runnable code. ## Languages Python, JavaScript, Go, .NET, Java, Rust, plus the Deepgram CLI (`dg` / `deepctl`). ## Install the related SDK skills For language-idiomatic patterns beyond a single recipe (full quick-starts, common patterns, gotchas), install the SDK-specific skills: ```bash npx skills add deepgram/deepgram-python-sdk # Python npx skills add deepgram/deepgram-js-sdk # JavaScript / TypeScript npx skills add deepgram/deepgram-java-sdk # Java npx skills add deepgram/deepgram-go-sdk # Go npx skills add deepgram/deepgram-rust-sdk # Rust npx skills add deepgram/deepgram-dotnet-sdk # C# / .NET ``` Swift and Kotlin SDK skills are not listed because those repositories are not public and `npx skills add` cannot reach them. For browser work, open the `browser-agent` skill: it covers the four Browser Agent SDK packages published on npm (`@deepgram/agents`, `@deepgram/react`, `@deepgram/ui`, `@deepgram/agents-widget`). ## Related Deepgram skills - `api` — consolidated REST + WebSocket API reference - `examples` — third-party platform integrations (Twilio, LiveKit, LangChain, etc.) - `starters` — runnable starter apps (framework × feature matrix) - `docs` — documentation finder - `setup-mcp` — Deepgram MCP server installation