WebRouter

WebRouter is Nimbleway's open-source Python library that gives an agent one call shape for web search and page reading across many providers. Its Tavily provider takes an api_base_url, so setting it to SpecStream's Tavily address sends WebRouter's search and extract calls to SpecStream. Your agent code stays as it is; only the router's config changes.

SpecStream for WebRouterAgent frameworks & builders

What you need

Python 3.11, 3.12 or 3.13, and WebRouter with its search and extract extras, which install Tavily's Python client: pip install "webrouter[search,extract]".
A SpecStream account. New accounts get $10 of credit, no card. Create one.

1
1. Create an API key
In your SpecStream workspace, open Agents & API, choose API keys, then Create API key, and set a total limit. Copy the key from the next screen. SpecStream shows it once.
Keep it in an environment variable where your agent runs:
Terminal
export SPECSTREAM_API_KEY="YOUR_SPECSTREAM_API_KEY"

2. Point the Tavily provider at SpecStream

provider_config is where WebRouter takes credentials and client settings, keyed by provider name. The Tavily provider hands api_base_url to Tavily's own Python clients, for search and for extract alike. Set it to SpecStream's Tavily address and make tavily the router's default provider:

router.py
import os

from web_router import WebRouter

router = WebRouter(
    "tavily",
    provider_config={
        "tavily": {
            "api_key": os.environ["SPECSTREAM_API_KEY"],
            "api_base_url": "https://api.specstream.ai/tavily",
        }
    },
)

response = router.search("what is retrieval augmented generation", max_results=5)
for result in response.results:
    print(result.title, result.url)

pages = router.extract(response.urls[:3], format="markdown")
for document in pages.documents:
    print(document.url, len(document.content))

3
3. Run it
Run python router.py. It prints the title and address of each result, then the address and length of each page it read. response.provider reports tavily, the provider class WebRouter used, while the calls went to api.specstream.ai.
To see the calls, open the Search API screen of your workspace. Each search is listed with its query and price.

What WebRouter sends and gets back

WebRouter calls Tavily's Python client, and the client posts to the base address you set:

WebRouter	SpecStream
`router.search(query, max_results=5)`	`POST /tavily/search` with `query` and `max_results`
`response.results[]`	`results[].title`, `url` and `content`, in Tavily's shape
`router.extract(urls, format="markdown")`	`POST /tavily/extract`, up to 20 pages a call
`pages.documents[].content`	`results[].raw_content`

With auto and other providers

Under auto, WebRouter's default, a backend given an api_key in provider_config joins the chain beside the backends whose keys are in your environment, then the keyless ones. With tavily named as the default, as above, every search and extract goes to SpecStream. To send one call elsewhere, pass provider= on that call.

AsyncWebRouter takes the same arguments: open it with async with AsyncWebRouter("tavily", provider_config=config) as router: and await router.search(...) and router.extract(...).

Cost

Each search is one fast search through the Tavily address: a fifth of the default search's $5 per 1,000 searches of up to 10 results. A max_results above 10 costs more in the same proportion, and a search that returns no results costs nothing. WebRouter's Tavily search sends no include_answer, so it stays on the fast tier.

Each page extract reads is $0.80 per 1,000 pages, and a page with no text is not charged. Both come out of your balance, with no quote and no approval.

For scale, WebRouter's own price list, PRICING.md, checked October 7, 2026, puts a default Tavily search at one credit of $0.008 and a default Nimble search at $5.00 per 1,000.

New accounts start with $10 of credit, no card. Balance and billing has the details.

Remove

Delete the tavily entry from provider_config, or drop its api_base_url. To stop the key working everywhere, revoke it in Agents & API. The same Tavily address serves other tools: see Moving off Tavily. Other ways to build with SpecStream are in agent frameworks. Questions go to [email protected].

SpecStream for WebRouter covers what it does in WebRouter on one page. Every other tool has its guide in Connect your tools. You see the most a request can cost before it starts, and the matches are paid from your balance. Billing has the details. For help, email [email protected].