@livekit/agents@1.8.1

# Duplex Model Support

Speech models that are able to speak and listen simultaneously are now supported via the new `DuplexModel` class, with OpenAI's GPT-Live model as the first to be implemented.

```typescript
import { voice } from '@livekit/agents';
import * as openai from '@livekit/agents-plugin-openai';

const session = new voice.AgentSession({
  llm: new openai.realtime.GPTLiveModel({
    voice: 'marin',
    // backend Responses model that handles the reasoning and the tool calls
    responsesOptions: {
      model: 'gpt-5.6-luna',
      instructions: 'Use tools when current information is required.',
    },
  }),
});
```

Read more about duplex models in [our docs](https://docs.livekit.io/agents/models/realtime/#full-duplex).

## What's Changed
* fix(telemetry): describe option objects in the session report by @rosetta-livekit-bot[bot] in https://github.com/livekit/agents-js/pull/2431
* fix: drop unused audio forwarding buffer by @rosetta-livekit-bot[bot] in https://github.com/livekit/agents-js/pull/2433
* plugins/phonic: add enableWatermarking realtime config option by @arunwpm-work in https://github.com/livekit/agents-js/pull/2424
* fix: name VAD in the slow-inference warning by @claude[bot] in https://github.com/livekit/agents-js/pull/2453
* fix: stop force interrupt inside a tool from crashing the process by @chenghao-mou in https://github.com/livekit/agents-js/pull/2444
* chore: enable type-aware promise linting by @chenghao-mou in https://github.com/livekit/agents-js/pull/2446
* fix: settle amd when session shutdown starts by @chenghao-mou in https://github.com/livekit/agents-js/pull/2451
* plugins/phonic: make allow_tool_chaining configurable per tool by @arunwpm-work in https://github.com/livekit/agents-js/pull/2466
* fix: serialize AgentConfigUpdate.instructions as a string in toJSON by @chenghao-mou in https://github.com/livekit/agents-js/pull/2467
* fix: preserve empty transcripts on interruption by @chenghao-mou in https://github.com/livekit/agents-js/pull/2464
* fix(amd): resolve transcript-only greetings at eot by @rosetta-livekit-bot[bot] in https://github.com/livekit/agents-js/pull/2400
* feat(meta): add Muse Voice streaming STT plugin by @rosetta-livekit-bot[bot] in https://github.com/livekit/agents-js/pull/2455
* feat(assemblyai): add universal-3-6-pro streaming model by @rosetta-livekit-bot[bot] in https://github.com/livekit/agents-js/pull/2409
* fix(assemblyai): validate live options and preserve Unicode context by @meeranmalik in https://github.com/livekit/agents-js/pull/2476
* fix: preserve final inference STT transcripts by @chenghao-mou in https://github.com/livekit/agents-js/pull/2465
* fix: wait for agent task handoffs before closing sessions by @chenghao-mou in https://github.com/livekit/agents-js/pull/2472
* feat: support full-duplex speech models by @chenghao-mou in https://github.com/livekit/agents-js/pull/2471
* feat: add the gpt-live duplex plugin by @chenghao-mou in https://github.com/livekit/agents-js/pull/2474
* Version Packages by @github-actions[bot] in https://github.com/livekit/agents-js/pull/2432

## New Contributors
* @claude[bot] made their first contribution in https://github.com/livekit/agents-js/pull/2453
* @meeranmalik made their first contribution in https://github.com/livekit/agents-js/pull/2476

**Full Changelog**: https://github.com/livekit/agents-js/compare/@livekit/agents-plugin-anam@1.8.0...@livekit/agents@1.8.1