# Deepgram SDK review

> Deepgram SDK is the official Python SDK for Deepgram's speech recognition, text-to-speech, text analysis, and voice agent APIs.

Devtoolwatch Board, the week of Oct 25–31, 2026. Canonical page: https://index.devtoolwatch.com/tools/deepgram-sdk/

## TL;DR

Deepgram SDK is the official Python client for Deepgram's speech-to-text, text-to-speech, language understanding, and voice agent APIs. Deepgram SDK suits Python developers building transcription, speech synthesis, or conversational voice apps. Deepgram SDK provides sync and async clients, WebSocket streaming, and pluggable transports, including a SageMaker transport.

## Deepgram SDK at a glance

- **Category:** [Voice and realtime](https://index.devtoolwatch.com/categories/voice-and-realtime/)
- **Maintained by:** Deepgram (organization, [@deepgram](https://github.com/deepgram) on GitHub)
- **License:** MIT
- **Language:** Python
- **Repository:** [github.com/deepgram/deepgram-python-sdk](https://github.com/deepgram/deepgram-python-sdk)
- **Website:** [developers.deepgram.com](https://developers.deepgram.com)
- **npm package:** [@deepgram/sdk](https://www.npmjs.com/package/@deepgram/sdk) 5.15.0
- **PyPI package:** [deepgram-sdk](https://pypi.org/project/deepgram-sdk/) 7.14.0
- **Latest release:** [v7.14.0](https://github.com/deepgram/deepgram-python-sdk/releases/tag/v7.14.0) on GitHub, Oct 20, 2026
- **First published:** May 12, 2021

## What Deepgram SDK does

- Deepgram SDK streams real-time speech recognition over WebSockets, including the Flux model with contextual turn detection.
- Deepgram SDK transcribes pre-recorded audio files and URLs, with explicit language selection, language detection, or multi-language mode.
- Deepgram SDK generates speech audio from text and supports pronunciation overrides and Flux batch pause markers.
- Deepgram SDK analyzes text for sentiment, topics, intents, and summaries.
- Deepgram SDK builds voice agents that combine Deepgram listening, an LLM think provider such as OpenAI, and Deepgram speech output.
- Deepgram SDK reconfigures active Nova-3 and Flux streams without reconnecting, including keyterms and numerals.
- Deepgram SDK covers project management, API keys, usage analytics, and access token generation.

### How it works

Deepgram SDK is a Python library you install with pip and call through DeepgramClient or AsyncDeepgramClient. Deepgram SDK sends requests to Deepgram's APIs over HTTP and WebSockets, authenticating with an API key or access token. Deepgram SDK lets you replace the built-in websockets transport, for example with deepgram-sagemaker to run models on AWS SageMaker endpoints.

## How to install Deepgram SDK

**With pip**

```
pip install deepgram-sdk
```

**SageMaker transport**

```
pip install deepgram-sagemaker
```

### Requirements

- Python 3.10 or later, below 4.0.
- A Deepgram API key or access token.
- Python 3.12 or later for the deepgram-sagemaker transport, plus AWS credentials.

### First steps

1. Install Deepgram SDK with pip install deepgram-sdk.
2. Set the DEEPGRAM_API_KEY environment variable, or pass api_key when creating DeepgramClient.
3. Create a client and call client.listen.v1.media.transcribe_file with model="nova-3" to transcribe an audio file.

## Who Deepgram SDK is for

- Python developers adding speech-to-text or text-to-speech to applications.
- Teams building real-time conversational voice agents.
- Developers running Deepgram models on AWS SageMaker endpoints.

### Limitations

- Deepgram SDK calls Deepgram's hosted APIs, so it needs a Deepgram API key or access token unless you use the SageMaker transport.
- Deepgram SDK's SageMaker transport is async-only and requires AsyncDeepgramClient.
- Deepgram SDK's nova-3 file transcription defaults to English when no language is given.
- Deepgram SDK's TTS pronunciation controls are Early Access, and pauses work only with Flux batch.

## Usage data (the week of Oct 25–31, 2026)

- Board rank: #289 overall, #15 in Voice and realtime, momentum score 33
- Downloads that week: 2.2M (+10% vs 4-week average)
- New GitHub stars per week (last 12 weeks): Aug 9: 1; Aug 16: 2; Aug 23: 0; Aug 30: 3; Sep 6: 4; Sep 13: 0; Sep 20: 3; Sep 27: 3; Oct 4: 1; Oct 11: 0; Oct 18: 0; Oct 25: 3

## Deepgram SDK alternatives

| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | [faster-whisper](https://index.devtoolwatch.com/tools/faster-whisper/) | 26.4k | 4.4M |
| #2 | [Kokoro](https://index.devtoolwatch.com/tools/kokoro/) | 9.8k | 150.7k |
| #3 | [Chatterbox](https://index.devtoolwatch.com/tools/chatterbox/) | 27.5k | 73.2k |
| #4 | [VoiceMem](https://index.devtoolwatch.com/tools/xzf-thu-voicemem/) | 2.8k | – |
| #5 | [Pipecat](https://index.devtoolwatch.com/tools/pipecat/) | 17k | 348.5k |
| #6 | [WhisperX](https://index.devtoolwatch.com/tools/whisperx/) | 25k | 158.6k |

## Deepgram SDK FAQ

### Is Deepgram SDK free and open source?

Yes. Deepgram SDK's source code is on GitHub (deepgram/deepgram-python-sdk) under the MIT license, so you can use it for free.

### How do I install Deepgram SDK?

Run `pip install deepgram-sdk`. It needs Node.js >=18.0.0. The install section above has the other options from the project's README.

### Is Deepgram SDK still maintained?

Yes. The deepgram/deepgram-python-sdk repository was last updated on Oct 20, 2026, and the latest release, v7.14.0 on GitHub, came out on Oct 20, 2026.

### How popular is Deepgram SDK?

Deepgram SDK has 469 GitHub stars as of Nov 1, 2026. It gained 3 new stars in the week of Oct 25–31, 2026, against a 12-week average of 2 a week. Its main package had about 2.2M downloads that week, +10% against its own 4-week average. That made it #15 among voice and realtime AI tools on the Devtoolwatch Board.

### What are the best alternatives to Deepgram SDK?

Other voice and realtime AI tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: faster-whisper (#1), Kokoro (#2), Chatterbox (#3), VoiceMem (#4), Pipecat (#5). The alternatives table above has their numbers.

### How does Deepgram SDK authenticate?

Deepgram SDK uses an API key for server-side apps or an access token for client-side or scoped access. Deepgram SDK reads DEEPGRAM_TOKEN or DEEPGRAM_API_KEY from the environment, and explicit parameters take precedence.

### Does Deepgram SDK support async code?

Yes. Deepgram SDK includes AsyncDeepgramClient with async/await support for file transcription and WebSocket connections.

### Does Deepgram SDK retry failed requests?

Yes. Deepgram SDK retries 408, 429, and 5xx responses with exponential backoff. You can set max_retries and timeouts per request or set a timeout on the client.

### Can Deepgram SDK transcribe non-English audio?

Yes. Pass a supported language such as language="fr", use detect_language=True when one dominant language is unknown, or language="multi" for mixed-language audio.

### Is there a Deepgram SDK for JavaScript?

Yes. Deepgram publishes @deepgram/sdk on npm, which requires Node.js 18 or later. Deepgram's docs also list Go, .NET, and Java SDKs.

## Sources

1. [deepgram/deepgram-python-sdk on GitHub](https://github.com/deepgram/deepgram-python-sdk)
2. [Deepgram SDK README](https://github.com/deepgram/deepgram-python-sdk/blob/main/README.md)
3. [@deepgram/sdk on npm](https://www.npmjs.com/package/@deepgram/sdk)
4. [deepgram-sdk on PyPI](https://pypi.org/project/deepgram-sdk/)
5. [Deepgram SDK website](https://developers.deepgram.com)
6. [Release v7.14.0 (Oct 20, 2026)](https://github.com/deepgram/deepgram-python-sdk/releases/tag/v7.14.0)
7. [Release v7.13.0 (Oct 12, 2026)](https://github.com/deepgram/deepgram-python-sdk/releases/tag/v7.13.0)
8. [Release v7.12.0 (Oct 2, 2026)](https://github.com/deepgram/deepgram-python-sdk/releases/tag/v7.12.0)
9. [Devtoolwatch Board, the week of Oct 25–31, 2026](https://index.devtoolwatch.com/)