--- title: "Strands Agents with an OpenAI-compatible model provider" canonical: https://uplinkswitch.ai/guides/strands-agents generated: 2026-11-04T07:23:47.011Z source: https://uplinkswitch.ai/guides/strands-agents base_url: https://api.uplinkswitch.com/v1 --- [›uplinkswitch](https://uplinkswitch.ai/) [api.uplinkswitch.com/v1](https://uplinkswitch.ai/models)[Console](https://uplinkswitch.ai/console) [›uplinkswitch](https://uplinkswitch.ai/) # Strands Agents The SDK’s OpenAIModel with one base URL and one key, in Python and TypeScript. Claude, GPT and DeepSeek agents with tools. Checked 2026-10-19 ·[All guides](https://uplinkswitch.ai/guides)·[Models](https://uplinkswitch.ai/models) ## Python Strands ships an OpenAI provider, and UplinkSwitch speaks the OpenAI format for every model. Install the SDK with the `openai` extra, create a key in the [console](https://uplinkswitch.ai/console) and export it as `UPLINKSWITCH_API_KEY`. shell ``` pip install 'strands-agents[openai]' ``` agent.py ``` import os from strands import Agent, tool from strands.models.openai import OpenAIModel model = OpenAIModel( client_args={ "api_key": os.environ["UPLINKSWITCH_API_KEY"], "base_url": "https://api.uplinkswitch.com/v1", "default_headers": { "HTTP-Referer": "https://uplinkswitch.ai", "X-Title": "Strands Agents", }, }, model_id="claude-sonnet-5-5", params={"max_tokens": 1000}, ) @tool def word_count(text: str) -> int: """Count the words in a piece of text.""" return len(text.split()) agent = Agent(model=model, tools=[word_count]) agent("How many words are in 'the quick brown fox jumps'? Use the word_count tool.") ``` `base_url` stops at `/v1`, and `model_id` is the id from the [Models page](https://uplinkswitch.ai/models). `params` goes into every request, so `max_tokens` and `temperature` belong there.`client_args` goes to the OpenAI client, and its two custom headers name the app in our request log. Claude Sonnet 5.5, GPT-6.1 Sol and DeepSeek V4 Pro each called `word_count` and answered 5 in our run. ## TypeScript The TypeScript SDK takes the key as `apiKey` and the base URL inside `clientConfig`, which also carries the `defaultHeaders` that name the app in our request log. `api: 'chat'` keeps it on Chat Completions. shell ``` npm install @strands-agents/sdk openai zod ``` agent.ts ``` import { Agent, tool } from '@strands-agents/sdk'; import { OpenAIModel } from '@strands-agents/sdk/models/openai'; import { z } from 'zod'; const model = new OpenAIModel({ api: 'chat', apiKey: process.env.UPLINKSWITCH_API_KEY, clientConfig: { baseURL: 'https://api.uplinkswitch.com/v1', defaultHeaders: { 'HTTP-Referer': 'https://uplinkswitch.ai', 'X-Title': 'Strands Agents', }, }, modelId: 'gpt-6.1-sol', }); const wordCount = tool({ name: 'word_count', description: 'Count the words in a piece of text.', inputSchema: z.object({ text: z.string() }), callback: (input) => input.text.split(/\s+/).length, }); const agent = new Agent({ model, tools: [wordCount] }); const result = await agent.invoke("How many words are in 'the quick brown fox jumps'?"); console.log(String(result)); ``` ## Structured output Pass a Pydantic model as `structured_output_model` on the call. Strands turns it into a tool, and the validated object lands on `result.structured_output`. structured.py ``` from pydantic import BaseModel class Capital(BaseModel): country: str capital: str result = Agent(model=model)("What is the capital of France?", structured_output_model=Capital) print(result.structured_output) # country='France' capital='Paris' ``` All three models returned the object this way. The deprecated `Agent.structured_output()` sends `response_format` instead; with it, Claude Sonnet 5.5 and DeepSeek V4 Pro answered in plain text. ## Streaming `stream_async` yields text events as the model writes. With the default callback handler, a plain `agent(...)` call prints the same stream to the terminal. stream.py ``` import asyncio async def main(): agent = Agent(model=model, callback_handler=None) async for event in agent.stream_async("Count from 1 to 5."): if "data" in event: print(event["data"], end="", flush=True) asyncio.run(main()) ``` ## Models and prices | Model | Model id | Input / 1M | Output / 1M | Developer list, in / out | | --- | --- | --- | --- | --- | | [Claude Sonnet 5.5](https://uplinkswitch.ai/models/claude-sonnet-5-5) | `claude-sonnet-5-5` | $1.20 | $3.00 | $2.00 / $10.00 | | [GPT-6.1 Sol](https://uplinkswitch.ai/models/gpt-6.1-sol) | `gpt-6.1-sol` | $0.60 | $3.00 | $2.00 / $10.00 | | [DeepSeek V4 Pro 0813](https://uplinkswitch.ai/models/deepseek-v4-pro-0813) | `deepseek-v4-pro-0813` | $0.792 | $1.188 | $1.32 / $3.96 | | [Claude Opus 5.5](https://uplinkswitch.ai/models/claude-opus-5-5) | `claude-opus-5-5` | $2.80 | $10.00 | $4.00 / $20.00 | | [Gemini 3.1 Pro Preview](https://uplinkswitch.ai/models/gemini-3.1-pro-preview) | `gemini-3.1-pro-preview` | $1.20 | $7.20 | $2.00 / $12.00 | | [Kimi K3](https://uplinkswitch.ai/models/kimi-k3) | `kimi-k3` | $0.90 | $4.50 | $3.00 / $15.00 | USD per million tokens, live from GET /v1/catalog. An agent sends the whole conversation on every model call, so input tokens add up fast. A task of five calls that together send 20,000 input tokens and get 1,500 back costs $0.0285 on Claude Sonnet 5.5, against $0.055 at Anthropic's list price, and $0.0176 on DeepSeek V4 Pro 0813. Credit is prepaid from $10 with no subscription. ## Check it Confirm the key before you start an agent: shell ``` curl https://api.uplinkswitch.com/v1/models \ -H "Authorization: Bearer $UPLINKSWITCH_API_KEY" \ -H "HTTP-Referer: https://uplinkswitch.ai" \ -H "X-Title: Strands Agents" ``` A list of model ids means the key works. The snippets on this page ran against the live API on 2026-10-19: the Python agent with its tool on Claude Sonnet 5.5, GPT-6.1 Sol and DeepSeek V4 Pro with strands-agents 1.61.1, the TypeScript agent on GPT-6.1 Sol with @strands-agents/sdk 1.21.0, plus streaming and structured output. ## Errors - A 401 with `The API key is not valid.` means the key is wrong or revoked; with no key the message is `No API key was sent.` - A 404 with `No such route.` means `base_url` lost its `/v1`. - A 404 with `No such model` means the `model_id` is not in the catalog. Copy it from the [Models page](https://uplinkswitch.ai/models). - A 402 with the code `insufficient_balance` means the prepaid balance ran out. Requests work again after a top-up. ## Questions ### Yes. The SDK's OpenAIModel provider passes client\_args to the OpenAI client, so a base\_url of https://api.uplinkswitch.com/v1 and an UplinkSwitch key are all it needs. In TypeScript the same values go in clientConfig and apiKey. ### Yes. Through OpenAIModel, claude-sonnet-5-5 and claude-opus-5-5 answer in the OpenAI format, tool calls included. In our run Claude Sonnet 5.5 called the word\_count tool and answered 5. ### strands-agents 1.61.1 for Python and @strands-agents/sdk 1.21.0 for TypeScript, on 2026-10-19. ### structured\_output\_model, passed to the agent call, returned the object on Claude Sonnet 5.5, GPT-6.1 Sol and DeepSeek V4 Pro. The older Agent.structured\_output() method is deprecated in Strands and returned the object only on GPT-6.1 Sol in our run. ### Each model call is a separate request, and the Usage page of the console lists every one with its token counts and charge. ## Related - [SDK setup](https://uplinkswitch.ai/guides/sdks) - [LangChain setup](https://uplinkswitch.ai/guides/langchain) - [Vercel AI SDK setup](https://uplinkswitch.ai/guides/vercel-ai-sdk) - [Claude Sonnet 5.5](https://uplinkswitch.ai/models/claude-sonnet-5-5) - [GPT-6.1 Sol](https://uplinkswitch.ai/models/gpt-6.1-sol) - [DeepSeek V4 Pro](https://uplinkswitch.ai/models/deepseek-v4-pro-0813) - [Model catalog](https://uplinkswitch.ai/models) - [LLM API price index](https://uplinkswitch.ai/pricing-index)