Claude for Legal adds a gallery of 19 ready-made tools, from cite checks to horizon scans

Anthropic has added a gallery of ready-made artifacts to Claude for Legal, and its announcement names contract review, regulatory horizon scans and cite checks among them. Per Anthropic’s page, the gallery has 19 artifacts, and teams can share them with clients.

Six Anthropic tools, four from partners, nine for access to justice

Anthropic’s own pitch is short: “Artifacts are interactive tools, such as trackers and dashboards, that teams can use and share in Claude.” In the gallery, each one is built for a specific legal task. The post on X says users can start from the built-in example, add their own documents and connectors, and then share the result.

Anthropic’s announcement page groups the artifacts in three sets:

• Six task artifacts built by Anthropic: cite check, jurisdiction survey matrix, matter board, playbook deviation board, regulatory horizon scan and tabular review.
• Four built by partners: an AI Governance Benchmark from Goodwin, Harvey Intake from Harvey, an Obligations Checker from Foley & Lardner, and a Subpoena Response Tool from PwC.
• Nine access-to-justice artifacts: the page names five, among them a Criminal Charge Reality Check from Zealous, an Immigration Papers Reader from Mobile Pathways, a Rent Ledger Reviewer and a Record-Sealing Screener from Rising for Justice, and Record Clearance from The Access Project. It does not name the other four.

Anthropic says more artifacts are coming from Legora and Orrick. Legal IT Insider, which also covered the launch, counted 12 access-to-justice artifacts rather than nine, so the exact total differs between reports. Legal IT Insider described the tabular review as answering questions across contracts and showing the supporting quote and page.

Enterprise only on paper, with no price listed

The announcement is filed under Claude Enterprise. It gives no pricing and no tier details beyond that, and it does not say whether Team customers get the gallery. Unite.ai, which reported the launch the same day, said customer data on Team and Enterprise plans is not used for training by default, citing Anthropic’s legal solutions page. That detail comes from a secondary report and was not on the announcement page itself.

The gallery sits at claude.ai/artifacts?templates=legal, and the Claude for Legal page is at claude.com/solutions/legal. Anthropic says the product connects to more than 60 legal connectors, including Bloomberg Law, Filevine, Ivo, Legora and Patlytics, with the full list in a legal directory at claude.ai/directory/legal. Company-wide rollouts go through Anthropic’s sales team or implementation partners: Accenture, AnswerRocket, Caylent, Deloitte, KPMG, Lazer, Perficient, phData, Protiviti, Provectus, PwC, Sia Partners and Tenex.

Sharing is the part aimed at law firms and legal departments. Artifacts can go to colleagues across a business and to clients and partners, and Anthropic says the settings are meant to protect privacy and support ethical walls between teams.

What would show this is more than a template shelf

Claude for Legal started in May. According to Unite.ai, that launch came with 12 practice-area plugins and more than 20 connectors, so the connector count has roughly tripled in about five months. Legal IT Insider reports that Anthropic’s own legal team already uses artifacts for contract redlining and playbook generation, privacy impact assessments, patent processes, marketing legal reviews and corporate legal attestations. Both figures are the company’s or reporters’ accounts of it, not independent measurements.

Several things would confirm or undercut the pitch. The first is whether the partner artifacts keep growing: Goodwin, Harvey, Foley & Lardner and PwC are in, Legora and Orrick are promised, and the gallery is only as useful as the firms that fill it. Harvey and Legora are also legal AI vendors in their own right, so their presence inside Claude is a signal to watch for how the market divides between general assistants and specialist products.

The second is accuracy. A cite check artifact is a direct test of the hallucinated-citation problem that has dogged legal AI. The announcement does not publish error rates or benchmarks, and Anthropic’s description of the tools as a starting point for lawyer review is a claim about workflow, not a measured result. Reviewers will want to see how often the cite check catches bad references, and how clearly each output points back to its source.

The third is access. Anthropic says it is funding law schools and public-interest groups through Claude credits, and Unite.ai reports a Legal Ambassadors Program waitlist and a 12-month Oxford programme backed by Microsoft that supports up to 10 innovators. Whether the nine access-to-justice artifacts reach people outside the pilots is not established. One of them, the Criminal Charge Reality Check, is listed by Unite.ai as covering New York only, and the rent and record-sealing tools reflect D.C. law, so jurisdiction limits are built in.

What the source does not establish: pricing, which plans beyond Enterprise can open the gallery, usage numbers, and any independent testing of the tools. Until those appear, the gallery is best read as a packaging move: Anthropic is turning a general assistant into a set of task-specific interfaces that lawyers can adapt and hand to clients.