Claude

Chat

Anthropic

Product overview

Name of Agent: Claude
Short description of agent: "Meet Claude, your AI problem solver and thinking partner. Claude works with you to write, research, code, and tackle complex problems with depth and precision. It enhances your thinking and expands what you're capable of" (for mobile apps, archived)
Date of release: 14/03/2023 30/05/2024 for tools 15/10/2025 latest (link, archived)
Advertised use: Learn/study, "Gather information from everywhere (search the web, connect Google Workspace, use custom integrations) and create interactive reports with reliable citations in artifacts (link, archived)." coding, research, " Analyze data, create visualizations, review documents, evaluate decisions, or break down key choi (link, archived)ces."
Monetisation/Usage price: Free, Pro (20), access to google workspace integrators and remote MCP Max (100), just higher rate limits
Who is using it?: end user customer, API usage, enterprises, government
Category: Chat

Company & accountability

Developer: Anthropic
Name of legal entity: Anthropic, PBC (link, archived)
Place of legal incorporation: Delaware, USA
For profit company?: Yes (PBC)
Parent company?: Not applicable
Governance documents analysis: Customer TOS (link, archived), Usage Policy (link, archived)
AI safety/trust framework: Responsible Scaling Policy (link, archived)
Compliance with existing standards: HIPAA, SOC 2 SOC 2 Type I and II, ISO 27001:2022 ISO/IEC 42001:2023 FedRAMP High, UK Cyber Essentials. (link, archived)

Technical capabilities & system architecture

Model specifications: Series of models, of which you select one. Reasoning is available. Details of available models can be found (here, archived).
Observation space: User inputted text and images, and an extensive system of tools, including internet access.
Action space: Text return to user, tool calls. (link, archived)
Memory architecture: Memory tool (link, archived)
User interface and interaction design: Chatbot. Also capable of generating artifacts for writing/code
User roles: Operator (issues queries, which the agent to responds to); Executor (user may take actions/make decisions based on outputs); Examiner (user can use thumbs up/down buttons to give feedback)
Component accessibility: Closed source

Autonomy & control

Autonomy level and planning depth: L1-L2: Tasks that users assigns to the agent are often narrow in scope. More complex tasks need multi-turn conversations where the user is in charge of planning. Agent always comes back to the user and awaits further instructions
User approval requirements for different decision types: The turn-based interaction paradigm by default requires user approval (issuing further instructions) to continue the interaction. Model can also ask follow-up/clarifying questions
Execution monitoring, traces, and transparency: Visible (albeit summarized) CoT when reasoning is activated
Emergency stop and shut down mechanisms and user control: User can pause/stop the agent at any time
Usage monitoring and statistics and patterns: None, will just tell you when you've hit their limit

Ecosystem interaction

Identify to humans?: Anthropic’s stance on watermarking (link, archived): "While watermarking is most commonly applied to image outputs, which we do not currently provide, we continue to work across industry and academia to explore and stay abreast of technological developments in this area." Anthropic’s Usage Policy prohibits using Claude to impersonate a human (i.e., to convince someone they’re communicating with a natural person when they are not), implying Claude deployments must not hide AI identity in human interactions (link, archived)
Identifies technically?: Anthropic officially documents that Claude-related web activity is identifiable via specific User-Agent tokens: ClaudeBot, Claude-User, and Claude-SearchBot (link, archived) Anthropic states it does not currently publish fixed IP ranges for these bots/agents (they use service-provider public IPs), so IP-range identification is not available as an official signature mechanism (link, archived)
Interoperability standards and integrations: Anthropic defines Model Context Protocol (MCP) as an open standard that “standardizes how applications provide context to LLMs,” likened to a “USB-C port for AI applications.” (link, archived)
Web conduct: - ClaudeBot, Claude-User, and Claude-SearchBot “respect ‘do not crawl’ signals by honoring industry standard directives in robots.txt” and “respect anti-circumvention technologies,” stating they do not attempt to bypass CAPTCHAs (link, archived) - Independent reporting and site-operator accounts, however, have documented periods of very heavy crawling and, at least in some cases, behavior that appeared to ignore site preferences until new robots.txt rules propagated. ((link, archived), (link, archived))

Safety, evaluation & impact

Technical guardrails and safety measures: Yes, Anthropic has a Usage Policy that they test/evaluate against. They also have internal fine-tuned Claude classifiers to catch harmful content. See (here, archived) for overview.
Sandboxing and containment approaches: Sandbox appears to be recommended but not required (see developer (docs, archived))
What types of risks were evaluated?: For Opus 4.5 (link, archived): CBRN, autonomy, cyber capabilities and misuse, child safety, bias
(Internal) safety evaluations and results: For Opus 4.5 (link, archived): - Child safety evaluations: exact results not disclosed - Political bias: Opus 4.5 scores 0.96 on political even-handness eval (link, archived) and 0.40 on acknowledging opposing perspectives eval - Bias/discrimination: on BBQ benchmark, Opus 4.5 has bias scores of -0.64 on disambiguated questions and 0.26 on ambiguous questions - CBRN: Anthropic "determined ASL-3 safeguards were appropriate...whereas Claude Opus 4.5 is our strongest biology model to date, we do not think it merits ASL-4 safeguards." - Autonomy: "Our determination is that Claude Opus 4.5 does not cross the AI R&D-4 capability threshold." - Cyber capabilities/misuse: "We believe that Claude Opus 4.5 does not demonstrate catastrophically risky capabilities in the cyber domain"
Third-party testing, audits, and red-teaming: For Opus 4.5 (link, archived): "pre-deployment testing of Claude Opus 4.5 was conducted by the US Center for AI Standards and Innovation (CAISI) and the UK AI Security Institute (UK AISI). These organizations conducted independent assessments focused on potential catastrophic risks in CBRN capabilities, cyber capabilities, ASL-3 safeguards, and misalignment." Anthropic also partners with the U.S. National Nuclear Security Administration for nuclear risks evaluation (results and data not disclosed).
Benchmark performance and demonstrated capabilities: Refer to Opus 4.5 System Card (link, archived)
Bug bounty programmes and vulnerability disclosure: Yes (link, archived)
Any known incidents?: None found
← Back to 2025 Index