deepgram · rohan-tessl · May 7, 2026 · Jun 3, 2026 · Jun 3, 2026
@@ -7,16 +7,19 @@ description: Use when writing or reviewing JavaScript/TypeScript in this repo th
 
 Analytics overlays applied to `/v1/listen`: summaries, topics, intents, sentiment, language detection, diarization, redaction, entities. Same client surface as STT; turn features on with parameters.
 
-## When to use this product
+**Use a different skill when:** plain transcription → `deepgram-js-speech-to-text`; analytics on text → `deepgram-js-text-intelligence`; Flux turn-taking → `deepgram-js-conversational-stt`; full-duplex agent → `deepgram-js-voice-agent`.
 
-- You have **audio** and want analytics returned alongside the transcript.
-- REST is the primary path; the WebSocket path supports only a subset of intelligence features.
+## Authentication
+
+```js
+require("dotenv").config();
 
-**Use a different skill when:**
-- You just want transcript output → `deepgram-js-speech-to-text`.
-- You already have text and want analytics on that text → `deepgram-js-text-intelligence`.
-- You need Flux turn-taking → `deepgram-js-conversational-stt`.
-- You need a full interactive voice agent → `deepgram-js-voice-agent`.
+const { DeepgramClient } = require("@deepgram/sdk");
+
+const deepgramClient = new DeepgramClient({
+  apiKey: process.env.DEEPGRAM_API_KEY,
+});
+```
 
 ## Feature availability: REST vs WSS
 
@@ -32,18 +35,6 @@ Analytics overlays applied to `/v1/listen`: summaries, topics, intents, sentimen
 | `sentiment` | yes | no |
 | `detect_language` | yes | no |
 
-## Authentication
-
-```js
-require("dotenv").config();
-
-const { DeepgramClient } = require("@deepgram/sdk");
-
-const deepgramClient = new DeepgramClient({
-  apiKey: process.env.DEEPGRAM_API_KEY,
-});
-```
-
 ## Quick start — REST with analytics
 
 From `examples/22-transcription-advanced-options.ts`:
@@ -70,6 +61,14 @@ const data = await deepgramClient.listen.v1.media.transcribeUrl({
   keyterm: ["keyword1", "keyword2"],
   redact: ["pci", "ssn"],
 });
+
+// Verify intelligence results are present
+const summary = data.results?.summary?.short;
+const topics = data.results?.topics?.segments;
+const sentiments = data.results?.sentiments?.segments;
+if (!summary && !topics && !sentiments) {
+  console.warn("No intelligence results — check feature/model/language support.");
+}
 ```
 
 ## Quick start — WSS subset
@@ -85,6 +84,13 @@ const deepgramConnection = await deepgramClient.listen.v1.createConnection({
 });
 ```
 
+## Workflow
+
+1. **Select features** from the REST vs WSS table. WSS lacks `summarize`, `topics`, `intents`, `sentiment`, `detect_language`.
+2. **Call** `transcribeUrl` / `transcribeFile` with chosen flags and `model: "nova-3"`.
+3. **Validate response**: check `data.results?.summary`, `data.results?.topics?.segments`, `data.results?.sentiments?.segments`. Fields are absent (not errored) when the model/language combo does not support the feature.
+4. **On missing results**: confirm the feature/model/language combination at https://developers.deepgram.com/docs/stt-intelligence-feature-overview, then retry with corrected params.
+
 ## Key parameters / API surface
 
 - Analytics flags: `summarize`, `topics`, `intents`, `sentiment`, `detect_language`, `detect_entities`, `diarize`, `redact`, `custom_topic`, `custom_topic_mode`, `custom_intent`, `custom_intent_mode`.
@@ -93,28 +99,18 @@ const deepgramConnection = await deepgramClient.listen.v1.createConnection({
 
 ## API reference (layered)
 
-1. **In-repo reference**: `reference.md` → `Listen V1 Media`; WSS subset behavior lives in `src/CustomClient.ts` and `src/api/resources/listen/resources/v1/client/{Client,Socket}.ts`.
-2. **Canonical OpenAPI (REST)**: https://developers.deepgram.com/openapi.yaml
-3. **Canonical AsyncAPI (WSS)**: https://developers.deepgram.com/asyncapi.yaml
-4. **Context7**: library ID `/llmstxt/developers_deepgram_llms_txt`
-5. **Product docs**:
-   - https://developers.deepgram.com/docs/stt-intelligence-feature-overview
-   - https://developers.deepgram.com/docs/summarization
-   - https://developers.deepgram.com/docs/topic-detection
-   - https://developers.deepgram.com/docs/intent-recognition
-   - https://developers.deepgram.com/docs/sentiment-analysis
-   - https://developers.deepgram.com/docs/language-detection
-   - https://developers.deepgram.com/docs/redaction
-   - https://developers.deepgram.com/docs/diarization
+1. **In-repo**: `reference.md` → `Listen V1 Media`; WSS subset in `src/api/resources/listen/resources/v1/client/{Client,Socket}.ts`.
+2. **OpenAPI / AsyncAPI**: https://developers.deepgram.com/openapi.yaml | https://developers.deepgram.com/asyncapi.yaml
+3. **Context7**: library ID `/llmstxt/developers_deepgram_llms_txt`
+4. **Product docs**: https://developers.deepgram.com/docs/stt-intelligence-feature-overview (links to summarization, topic detection, intent recognition, sentiment, language detection, redaction, diarization).
 
 ## Gotchas
 
-1. **`summarize` on `/v1/listen` is versioned, not plain boolean.** The generated REST surface and examples point at `"v2"`.
-2. **Most intelligence flags are REST-only.** Current WSS connect args do not expose `topics`, `intents`, `sentiment`, `summarize`, or `detect_language`.
-3. **`redact` typing is looser in practice than in the generated alias.** Examples pass arrays like `["pci", "ssn"]`, even though `ListenV1Redact` itself is just a string alias.
-4. **Use `keyterm` for Nova-3 biasing.** `examples/22-transcription-advanced-options.ts` explicitly notes keywords are not supported for Nova-3.
-5. **Model/feature support is product-side.** `nova-3` is the safest choice when mixing many overlays.
-6. **Diarization quality depends on audio quality and duration.** Short or noisy clips churn speakers.
+1. **`summarize` is `"v2"`, not boolean.** The generated REST surface and examples use the string value.
+2. **`redact` accepts arrays** like `["pci", "ssn"]` despite `ListenV1Redact` being a string alias.
+3. **Use `keyterm`, not `keywords`, for Nova-3 biasing.**
+4. **Prefer `nova-3`** when mixing many overlays -- broadest feature support.
+5. **Diarization quality depends on audio quality and duration.** Short or noisy clips churn speakers.
 
 ## Example files in this repo
 
@@ -125,10 +121,4 @@ const deepgramConnection = await deepgramClient.listen.v1.createConnection({
 
 ## 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:
-
-```bash
-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`).
+For cross-language Deepgram product knowledge, install `npx skills add deepgram/skills`.
@@ -7,18 +7,7 @@ description: Use when writing or reviewing JavaScript/TypeScript in this repo th
 
 Administrative REST endpoints under `/v1/projects`, `/v1/models`, and related project subresources.
 
-## When to use this product
-
-- **Projects**: list, get, update, delete, leave.
-- **Keys**: list, get, create, delete API keys.
-- **Members + invites**: inspect members, update scopes, create/delete invites.
-- **Usage + billing**: inspect requests, usage, usage breakdown, balances, purchases, billing breakdown.
-- **Models**: list global models and project-scoped models.
-- **Agent think models**: discover available model providers for Voice Agent `think` settings.
-
-**Use a different skill when:**
-- You want to run a live websocket agent session → `deepgram-js-voice-agent`.
-- You want transcription or synthesis calls rather than project/admin APIs → product-specific skills.
+**Use a different skill when:** live agent session → `deepgram-js-voice-agent`; transcription or synthesis → product-specific skills.
 
 ## Authentication
 
@@ -80,6 +69,13 @@ Think-model discovery for Voice Agent:
 await deepgramClient.agent.v1.settings.think.models.list();
 ```
 
+## Workflow for destructive operations
+
+1. **List** the resources first (e.g., `projects.keys.list(projectId)`).
+2. **Confirm** the target ID with the user before proceeding.
+3. **Execute** the delete/leave/remove call.
+4. **Verify** by listing again to confirm deletion.
+
 ## Key parameters / API surface
 
 - Projects: `client.manage.v1.projects.list/get/update/delete/leave`.
@@ -119,24 +115,8 @@ The current JS SDK does **not** expose persisted Voice Agent configuration CRUD
 
 ## Example files in this repo
 
-- `examples/13-management-projects.ts`
-- `examples/14-management-keys.ts`
-- `examples/15-management-members.ts`
-- `examples/16-management-invites.ts`
-- `examples/17-management-usage.ts`
-- `examples/18-management-billing.ts`
-- `examples/19-management-models.ts`
-- `examples/29-management-usage-breakdown.ts`
-- `examples/30-management-billing-detailed.ts`
-- `examples/31-management-member-permissions.ts`
-- `examples/32-management-project-models.ts`
+`examples/13-management-projects.ts` through `examples/19-management-models.ts`, plus `examples/29-32-*` for usage breakdown, billing details, member permissions, and project models.
 
 ## 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:
-
-```bash
-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`).
+For cross-language Deepgram product knowledge, install `npx skills add deepgram/skills`.
@@ -1,19 +1,13 @@
 ---
 name: deepgram-js-text-intelligence
-description: Use when writing or reviewing JavaScript/TypeScript in this repo that calls Deepgram Text Intelligence / Read (`/v1/read`) for sentiment, summarization, topic detection, and intent recognition on text input. Covers `client.read.v1.text.analyze(...)` with `body: { text }` or `body: { url }`. Use `deepgram-js-audio-intelligence` when the source is audio instead of text. Triggers include "read API", "text intelligence", "analyze text", "sentiment", "summarize text", "topics", "intents", and "read.v1".
+description: "Use when writing or reviewing JavaScript/TypeScript in this repo that calls Deepgram Text Intelligence / Read (`/v1/read`) for sentiment, summarization, topic detection, and intent recognition on text input. Covers `client.read.v1.text.analyze(...)` with `body: { text }` or `body: { url }`. Use `deepgram-js-audio-intelligence` when the source is audio instead of text. Triggers: read API, text intelligence, analyze text, sentiment, summarize text, topics, intents, read.v1."
 ---
 
 # Using Deepgram Text Intelligence (JavaScript / TypeScript SDK)
 
 Analyze text or a hosted text URL for sentiment, summarization, topics, and intents via `/v1/read`.
 
-## When to use this product
-
-- You already have **text** (transcript, document, email, chat log) and want analytics.
-- You want a single REST call; there is no streaming Read API in this SDK.
-
-**Use a different skill when:**
-- Your source is audio and you want the analytics applied during transcription → `deepgram-js-audio-intelligence`.
+**Use a different skill when:** source is audio → `deepgram-js-audio-intelligence`. This API is REST-only; there is no streaming Read API in this SDK.
 
 ## Authentication
 
@@ -85,10 +79,4 @@ For broader coverage, `examples/28-text-intelligence-advanced.ts` also demonstra
 
 ## 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:
-
-```bash
-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`).
+For cross-language Deepgram product knowledge, install `npx skills add deepgram/skills`.
@@ -7,13 +7,9 @@ description: Use when writing or reviewing JavaScript/TypeScript in this repo th
 
 Convert text to audio with one-shot REST generation or low-latency streaming synthesis via `/v1/speak`.
 
-## When to use this product
+Two modes: **REST** (`client.speak.v1.audio.generate`) for one-shot synthesis, **WebSocket** (`client.speak.v1.createConnection()`) for low-latency streaming.
 
-- **REST (`client.speak.v1.audio.generate`)** — render finished text into an audio response. Best for downloadable files, pre-generated prompts, batch synthesis.
-- **WebSocket (`client.speak.v1.createConnection()` / `connect()`)** — stream text in and receive audio out with lower latency. Best when an LLM is still producing tokens.
-
-**Use a different skill when:**
-- You need the agent to also listen, think, and handle barge-in → `deepgram-js-voice-agent`.
+**Use a different skill when:** full-duplex agent with STT + LLM + TTS → `deepgram-js-voice-agent`.
 
 ## Authentication
 
@@ -71,6 +67,8 @@ deepgramConnection.sendText({ type: "Speak", text: "Hello from streaming TTS." }
 deepgramConnection.sendFlush({ type: "Flush" });
 ```
 
+**Error handling:** Listen for `Warning` events in the message handler. If the connection drops, create a new connection and re-register handlers; the SDK does not auto-reconnect.
+
 ## Key parameters / API surface
 
 - REST & WSS: `model`, `encoding`, `sample_rate`, `container`, `bit_rate`, `callback`, `callback_method`, `tag`, `mip_opt_out`.
@@ -111,10 +109,4 @@ Unlike the Python SDK, this repo does **not** include a hand-written `TextBuilde
 
 ## 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:
-
-```bash
-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`).
+For cross-language Deepgram product knowledge, install `npx skills add deepgram/skills`.
@@ -7,16 +7,7 @@ description: Use when writing or reviewing JavaScript/TypeScript in this repo th
 
 Full-duplex voice agent runtime over `wss://agent.deepgram.com/v1/agent/converse`: audio in, LLM orchestration, audio out, plus function calling and prompt/runtime updates.
 
-## When to use this product
-
-- You want an **interactive voice assistant** where the user speaks, the agent thinks, and the agent responds with speech.
-- You need **function / tool calling** inside the conversation loop.
-- You want Deepgram to host the STT + think + TTS orchestration.
-
-**Use a different skill when:**
-- You only need transcription → `deepgram-js-speech-to-text` or `deepgram-js-conversational-stt`.
-- You only need synthesis → `deepgram-js-text-to-speech`.
-- You want project keys, usage, models, or other admin APIs → `deepgram-js-management-api`.
+**Use a different skill when:** transcription only → `deepgram-js-speech-to-text` or `deepgram-js-conversational-stt`; synthesis only → `deepgram-js-text-to-speech`; admin APIs → `deepgram-js-management-api`.
 
 ## Authentication
 
@@ -72,6 +63,17 @@ deepgramConnection.sendSettings({
 
 The same example also shows `client.agent.v1.settings.think.models.list()` for discovering supported think models.
 
+## Workflow
+
+1. `createConnection()` — returns a lazy socket; no network call yet.
+2. Register `on("message", ...)` handlers for `SettingsApplied`, `ConversationText`, `FunctionCallRequest`, `Error`, and audio payloads.
+3. `connect()` then `await waitForOpen()`.
+4. `sendSettings({ type: "Settings", ... })` — **must be the first message**. Wait for `SettingsApplied` before proceeding.
+5. `sendMedia(chunk)` to stream user audio. Send `sendKeepAlive(...)` every ~5 s during silence.
+6. Handle `FunctionCallRequest` with `sendFunctionCallResponse({ type: "FunctionCallResponse", id, name, content })`.
+7. Use `sendUpdatePrompt(...)`, `sendUpdateThink(...)`, `sendUpdateSpeak(...)` for runtime changes.
+8. On `Error` event, log the error and close/reconnect as appropriate.
+
 ## Key parameters / API surface
 
 - Connection setup: `client.agent.v1.createConnection()` / `connect()`.
@@ -114,10 +116,4 @@ This SDK exposes the **live agent runtime** plus `settings.think.models.list()`,
 
 ## 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:
-
-```bash
-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`).
+For cross-language Deepgram product knowledge, install `npx skills add deepgram/skills`.