LIVE
Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|
Toolofficial publisher

web_search (Anthropic)

Tool : A precise action the AI can call · requires anthropic-api

Server-side web search by Anthropic, with citations, enabled by a parameter in the API call.

The simplest way to give Claude access to fresh web content; the retrieved content may contain injected instructions.

Overall score
7.3 /10
Security: one caveat, see the card
Utility
9.0
Reliability
6.0
Security
One caveat
Ease
1 click · Comfortable

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

Claude, Anthropic's model, has a knowledge cutoff date and cannot answer questions about recent news, real-time prices, or recent events. The web_search tool solves this by giving Claude direct access to the web at the time of the request.

In practice, you simply add the web_search type parameter to your API call. Claude decides on its own when to run a search based on the question, the API executes the search server-side at Anthropic, then returns the results to Claude, which formulates its final response with citations to the sources. This cycle can repeat multiple times within a single call.

Three versions are available: the basic version (web_search_20250305), a version with dynamic filtering that reduces token consumption by keeping only relevant content (web_search_20260209), and a version with response inclusion control useful for agentic workflows (web_search_20260318). Dynamic filtering relies on server-side code execution, provisioned automatically at no extra charge.

This tool is aimed at developers integrating Claude via the API who want their applications to respond with up-to-date information. It is not available to end users without API access, and is not accessible on Amazon Bedrock or certain specific cloud configurations.

What it offers

  • ·Real-time server-side web search, triggered automatically by Claude based on the request (paid).
  • ·Systematic source citations in the final response, with URL, title, and excerpt of cited text (paid).
  • ·Dynamic filtering of results via code execution to reduce token consumption, available from version web_search_20260209 (paid).
  • ·max_uses parameter to limit the number of searches per request and control costs (paid).
  • ·Domain filtering via allowed_domains or blocked_domains to restrict or exclude certain sources (paid).
  • ·Localization of results using the user_location parameter (city, region, country, timezone) (paid).
  • ·Response inclusion control with response_inclusion to lighten responses in agentic workflows, available from version web_search_20260318 (paid).

In practice

Price
Paid · from 9 €
Commercial use
Commercial use: yes
Available on
api
Host
anthropic-api
FR interface
No

Security and data

Rating
Security: one caveat, see the card
Capabilities
readswritesexecutesnetwork
Your data
Data with the vendor
Maintenance
Active

Links and repository checked on 25 October 2026 · added on 1 May 2025

How to install web_search (Anthropic)

1. Obtain an Anthropic API key and enable billing on your account. 2. Check in the Claude Console that web search is enabled for your organization (an administrator can disable it). 3. Install the Anthropic Python SDK: pip install anthropic 4. Add the tools parameter with type web_search_20250305 (or a later version) to your client.messages.create() call. 5. Security note: web results are untrusted content; filter domains with allowed_domains or blocked_domains to limit exposure to unknown sources. 6. For multi-turn conversations, send back encrypted_content and encrypted_index blocks without modification.
pip install anthropic

Official documentation ↗

In the same category

Whole category Feed (memory, context) →