Baseten

OverviewCopy Anchor Link

Baseten's Model APIs serve open-weights models (GLM, Kimi, DeepSeek, Nemotron, gpt-oss) over an OpenAI-compatible API, and dedicated deployments run your own model on your own GPUs.

• BasetenLLMService: an OpenAILLMService subclass with Baseten's base URL and default model.

InstallationCopy Anchor Link

bun add @pipedog/baseten

It depends on @pipedog/openai, which is installed with it.

import { BasetenLLMService } from "@pipedog/baseten/llm";

PrerequisitesCopy Anchor Link

• A Baseten account and an API key from the Baseten console.
• Pass the key as apiKey. It is required, and the service never falls back to OPENAI_API_KEY. The examples read it from BASETEN_API_KEY.

BasetenLLMServiceCopy Anchor Link

ConfigurationCopy Anchor Link

The other options (organization, project, defaultHeaders, serviceTier, retryOnTimeout, retryTimeoutSecs, function calling) are OpenAILLMService's. See OpenAI.

SettingsCopy Anchor Link

Runtime-updatable settings, passed as settings: new BasetenLLMService.Settings({...}) (BasetenLLMSettings). They are OpenAI's settings, with these defaults:

Every other setting keeps OpenAI's default. Unset sampling settings are left out of the request, so the provider's own defaults apply. extra carries provider-specific request fields.

UsageCopy Anchor Link

import { LLMUpdateSettingsFrame } from "@pipedog/core/frames/frames";

const llm = new BasetenLLMService({
  apiKey: process.env.BASETEN_API_KEY ?? "",
  settings: new BasetenLLMService.Settings({
    model: "zai-org/GLM-5.3",
    systemInstruction: "You are a helpful assistant in a voice conversation.",
    temperature: 0.7,
  }),
});

A dedicated deployment: point baseUrl at its /sync/v1 URL and set model to the model it serves.

// Replace {model_id} with your deployment's ID.
const dedicated = new BasetenLLMService({
  apiKey: process.env.BASETEN_API_KEY ?? "",
  baseUrl: "https://model-{model_id}.api.baseten.co/environments/production/sync/v1",
  settings: new BasetenLLMService.Settings({ model: "Qwen/Qwen2.5-3B-Instruct" }),
});

Change settings mid-call with an LLMUpdateSettingsFrame. The next completion uses them:

// Queue it on the worker, or push it from a processor upstream of the service.
const update = new LLMUpdateSettingsFrame({
  delta: new BasetenLLMService.Settings({ temperature: 0.2, maxTokens: 300 }),
});

NotesCopy Anchor Link

• Runs a completion for each LLMContextFrame, between an LLMFullResponseStartFrame and an LLMFullResponseEndFrame, and streams the reply as LLMTextFrames, as OpenAILLMService does. Function calls, token usage metrics and runInference work the same way.
• "developer" messages are sent as "user" messages, since the API doesn't support that role (supportsDeveloperRole is false).
• Baseten's message roles are system, user, assistant and tool, so "developer" messages are sent as "user" messages.

Runtime notesCopy Anchor Link

Generated from pipecat.runtimes and pipecat.compat in packages/baseten/package.json: ✅ means the compat smoke test passes on that runtime.

• The client is the official openai JS SDK (through @pipedog/openai), which talks through fetch. It works on Bun, Node, Deno and Cloudflare Workers.
• The deprecated model constructor option is not ported. Use settings.
• Tested against a local OpenAI-compatible mock server, not against Baseten itself.