# Harvey

> Source: https://theagentsindex.com/harvey-ai

Legal AI platform whose agents run end-to-end legal work for law firms and in-house teams.

Harvey is a legal AI platform for law firms and in-house teams, pairing document vaults, legal research and contract review with agents that take a task end to end and return cited work. Admins run it as a governed enterprise workspace.

- Best for: Law firms and in-house legal teams running high-stakes matters on one governed platform

Vendor's price

On request

From [harvey.ai/contact-sales](https://www.harvey.ai/contact-sales), checked 6 November 2026

- MCP server Partly
- API Yes
- Open source Not stated
- Self-host No

What we checked

- Pages last read: 6 November 2026
- Facts checked: 20 facts From 19 sources, read 18 October 2026
- AI answers: Named in 1 of 151 Our October check

[How we check](https://theagentsindex.com/methodology) [Editorial policy](https://theagentsindex.com/editorial-policy#not-for-sale)

[Visit Harvey](https://theagentsindex.com/go/harvey-ai) [Compare](https://theagentsindex.com/compare?cat=legal-agents&products=harvey-ai) [Alternatives](https://theagentsindex.com/harvey-ai/alternatives)

## What we checked on Harvey

Yes Partly No

What it does No

### Done-for-you service

Harvey sells software a firm runs itself on one platform, not a consultancy engagement where Harvey's people do the legal work.

> Firms and legal departments do their best work with Harvey. Run everything your team and practice needs on one collaborative platform.

Homepage, 'One Platform for Legal Work' section intro [harvey.ai](https://www.harvey.ai/) Checked 18 October 2026

How much it does unattended Partly

### Runs autonomously

Agents execute tasks end to end and run in the background or on a schedule, but the vendor keeps a human approving the plan and reviewing output.

> Preview the plan, adjust the scope, and approve work before Harvey begins.

Agents page, 'Control the Quality of Work' - 'Review and Approve Work' [harvey.ai/platform/agents](https://www.harvey.ai/platform/agents) Checked 18 October 2026

How much it does unattended Yes

### Multi-agent

The vendor states that one task can be split across several agents working in parallel, with routing to the right agent.

> Hand off one task or ten. Harvey coordinates multiple agents to complete work in parallel and deliver review-ready outputs.

Agents page, 'Parallel Execution' feature block [harvey.ai/platform/agents](https://www.harvey.ai/platform/agents) Checked 18 October 2026

How much it does unattended Partly

### Agent permissions

Scope is approved before a run and the agent escalates to a human when judgment is needed; API tokens additionally carry per-endpoint permission lists.

> Harvey proactively brings you in when your judgment is needed, then keeps work moving.

Agents page, 'Nudges' feature block [harvey.ai/platform/agents](https://www.harvey.ai/platform/agents) Checked 18 October 2026

Which models it runs on Yes

### Claude

Anthropic models are integrated across the platform, with the vendor naming Claude Opus and Sonnet releases running in Harvey products.

> Claude Opus 4.6, for example, excels at agentic reasoning and deep multi-step legal analysis

Blog 'Why Harvey is Multi-Model by Design', 'Quality: The Best Model for Every Task' [harvey.ai/blog/why-harvey-is-multi-mode…](https://www.harvey.ai/blog/why-harvey-is-multi-model-by-design) Checked 18 October 2026

Which models it runs on Yes

### GPT

OpenAI models are among the frontier models incorporated, and the vendor describes GPT-5.2 behaviour inside the platform.

> while GPT-5.2 pairs strong analytical abilities with capability awareness

Blog 'Why Harvey is Multi-Model by Design', 'Quality: The Best Model for Every Task' [harvey.ai/blog/why-harvey-is-multi-mode…](https://www.harvey.ai/blog/why-harvey-is-multi-model-by-design) Checked 18 October 2026

Which models it runs on No

### Open models

The vendor's own account of its multi-model architecture names only three closed frontier providers, with no open-weight option offered to customers.

> expanding its platform to incorporate leading foundation models from Anthropic, Google DeepMind, and OpenAI

Blog 'Why Harvey is Multi-Model by Design', opening paragraph [harvey.ai/blog/why-harvey-is-multi-mode…](https://www.harvey.ai/blog/why-harvey-is-multi-model-by-design) Checked 18 October 2026

Which models it runs on Yes

### Model choice

Workspace admins decide which providers are enabled and individual users can pick a model in the selector, so the choice is the customer's.

> Harvey allows workspace administrators to determine which models are available to their organization.

Blog 'Why Harvey is Multi-Model by Design', 'Choice: Putting Control Where it Belongs' [harvey.ai/blog/why-harvey-is-multi-mode…](https://www.harvey.ai/blog/why-harvey-is-multi-model-by-design) Checked 18 October 2026

Where you use it Yes

### In a browser

Harvey is a hosted workspace reached in the browser, with admin settings pages, plus Word and Outlook add-ins and native mobile apps alongside it.

> To create and manage API tokens for your organization’s workspace, visit the "API Tokens" section in your workspace settings.

Authentication page, info callout at top [developers.harvey.ai/api-reference/auth…](https://developers.harvey.ai/api-reference/authentication) Checked 18 October 2026

Whose machine it runs on No

### Self-hosted

Harvey runs only in its own Azure cloud environment; the customer choice offered is which region processes the data, not which infrastructure.

> Harvey hosts its cloud environment in Microsoft Azure.

Security page FAQ, 'Where is my data hosted and processed?' [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

What it costs to run Yes

### API access

A documented REST API with bearer tokens covers assistant completions, Vault projects, review tables, audit logs and client matters.

> The Harvey API uses bearer token authentication. Include the token in the Authorization header of all API requests.

Authentication page, first body paragraph [developers.harvey.ai/api-reference/auth…](https://developers.harvey.ai/api-reference/authentication) Checked 18 October 2026

What it costs to run Partly

### MCP server

A hosted MCP server exposes five Harvey tools to Claude, Gemini and Copilot, but the docs say a user must be enrolled in the feature first.

> The Harvey MCP Server exposes some of Harvey's core features — legal Q&A, Vault document analysis, and research knowledge sources as MCP tools.

Harvey MCP guide, 'What is the Harvey MCP Server?' [developers.harvey.ai/guides/harvey_mcp.…](https://developers.harvey.ai/guides/harvey_mcp.md) Checked 18 October 2026

Buying it for a team Yes

### A company can buy it

The unit of sale is an organisation workspace with admin settings, named customer success contacts and firm-wide deployments.

> Rate limits are applied per organization and are reset every minute.

Rate Limits page, opening paragraph [developers.harvey.ai/api-reference/rate…](https://developers.harvey.ai/api-reference/ratelimits.md) Checked 18 October 2026

Buying it for a team Yes

### Admin controls

Admins constrain the agent and its users through role-based access, model availability, IP allow-listing, retention rules and synced ethical walls.

> Harvey enforces strict role-based access controls and logical workspace separation to ensure that only authorized users can access specific customer…

Security page FAQ, 'How do you respect access controls for client data?' [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

Buying it for a team Yes

### Audit log

Audit logs are both an enterprise control and a documented API, queryable and paginable for workspace activity monitoring.

> Audit Logs: Query and retrieve audit logs for workspace activity monitoring

Documentation index, Guides list entry for Audit Logs [developers.harvey.ai/llms.txt](https://developers.harvey.ai/llms.txt) Checked 18 October 2026

Buying it for a team Yes

### Single sign-on

SAML SSO is named as a default enterprise control alongside audit logs and IP allow-listing; OIDC is not mentioned.

> delivering 24/7 monitoring alongside enterprise-grade controls, including SAML SSO, audit logs, IP allow-listing, and data lifecycle management

Security page, 'Purpose-Built Security' card [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

What happens to your code Yes

### Opt out of training

No training on customer data is the contractual default rather than a setting, and the vendor requires zero data retention from model providers.

> We don’t use inputs, outputs, or uploaded documents to train underlying models.

Security page, 'No Model Training' card [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

What happens to your code Yes

### Data residency

Processing can be kept in-region in the EU or Switzerland, the US or Australia, and the regional API endpoints match those deployments.

> set retention policies, delete data anytime, and keep everything in-region—EU or Switzerland, US, or Australia

Security page, 'Data Sovereignty and Control' card [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

What happens to your code Partly

### Take your data out

Customers set retention and can delete data at any time, and usage and query history are exportable by API, but no document bulk-export or credit-expiry terms are published.

> Customers can determine what data to upload to Harvey, how long it is retained, and whether that data can be shared with others within the company.

Security page FAQ, 'How do you respect access controls for client data?' [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

What happens to your code Yes

### Certifications

Annually audited SOC 2 Type II and ISO 27001, with ISO 27701 and 42001, GDPR and CCPA also claimed and third-party penetration testing.

> Harvey undergoes annual SOC 2 Type II and ISO 27001 audits to validate these controls.

Security page FAQ, 'How does Harvey keep my data private and secure?' [harvey.ai/security](https://www.harvey.ai/security) Checked 18 October 2026

## For AI agents

- MCP server: Partly
- API: Yes

## Harvey alternatives

[All 3 alternatives](https://theagentsindex.com/harvey-ai/alternatives)

- [Spellbook](https://theagentsindex.com/spellbook) · AI contract review, drafting and autonomous contract management inside Microsoft Word. · On quote · Checked 8 November 2026 · [Visit](https://theagentsindex.com/go/spellbook)
  - Price: On quote
- [GC AI](https://theagentsindex.com/gc-ai) · Legal AI platform for in-house counsel, spanning drafting, contract review, playbooks and research. · Starts at $500 a seat a month · Checked 5 November 2026 · [Visit](https://theagentsindex.com/go/gc-ai)
  - From: $500 a seat a month
- 1 more alternative · 1 of the 3 publish a price. · [See all 3](https://theagentsindex.com/harvey-ai/alternatives)
  - [Legora](https://theagentsindex.com/legora) On quote Agentic operating system where law-firm and in-house teams run legal work end-to-end with AI agents.

## Questions about Harvey

- How much does Harvey cost?: Harvey gives its price on request. From harvey.ai/contact-sales, checked 6 November 2026.
- Does Harvey have an API or an MCP server?: It has an API. It has a limited MCP server.
- Is Harvey open source, and can you host it yourself?: You cannot host it yourself.
- What have we checked on Harvey?: 20 facts, under these headings: What it does, how much it does unattended, which models it runs on, where you use it, whose machine it runs on, what it costs to run, buying it for a team, what happens to your code.
- How do we check Harvey?: We hold 20 facts on Harvey, quoted from 19 pages. Facts last checked 18 October 2026. Price last checked 6 November 2026.

[How we check every listing](https://theagentsindex.com/methodology)

## Sources and changes

Last changed 6 November 2026

Sources 19

- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/security) /security
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform) /platform
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/agents) /platform/agents
- Read 18 October 2026 · [harvey.ai](https://www.harvey.ai/blog/why-harvey-is-multi-model-by-design) /blog/why-harvey-is-multi-model-by-design
- Read 18 October 2026 · [developers.harvey.ai](https://developers.harvey.ai/api-reference/authentication) /api-reference/authentication
- Read 18 October 2026 · [harvey.ai](https://www.harvey.ai/) /
- Read 18 October 2026 · [developers.harvey.ai](https://developers.harvey.ai/api-reference/ratelimits.md) /api-reference/ratelimits.md
- Read 18 October 2026 · [harvey.ai](https://www.harvey.ai/contact-sales) /contact-sales
- Read 18 October 2026 · [developers.harvey.ai](https://developers.harvey.ai/guides/harvey_mcp.md) /guides/harvey_mcp.md
- Read 18 October 2026 · [developers.harvey.ai](https://developers.harvey.ai/llms.txt) /llms.txt
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/command-center) /platform/command-center
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/connector-library) /platform/connector-library
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/contract-intelligence) /platform/contract-intelligence
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/facts-and-chronologies) /platform/facts-and-chronologies
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/harvey-mobile) /platform/harvey-mobile
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/horizon-scanning) /platform/horizon-scanning
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/intake) /platform/intake
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/memory) /platform/memory
- Read 6 November 2026 · [harvey.ai](https://www.harvey.ai/platform/vault) /platform/vault

## Similar to Harvey

- [Aissist](https://theagentsindex.com/aissist) Multi-agent AI for support and sales that works inside your helpdesk, billed per interaction with a per-resolution cap. Free plan, then $0.09 per interaction
- [Fin](https://theagentsindex.com/intercom-fin) Intercom's customer agent that resolves support conversations across chat, email, voice and social channels. From $0.99 per outcome
- [Zendesk AI agents](https://theagentsindex.com/zendesk-ai-agents) Customer-service AI agents for messaging, email and voice, billed per automated resolution. From about $28 a seat a month
- [Botpress](https://theagentsindex.com/botpress) Customer support AI agents and helpdesk, priced per conversation with AI usage included and no per-seat fees. Free plan, then $189 a month

## From our research

[All research](https://theagentsindex.com/data)

- [Data report](https://theagentsindex.com/blog/terminal-bench-coding-agent-scores-2026) 2 Aug 2026 Terminal-Bench 2.1: how our indexed coding agents actually score Stanford/Laude Institute's independent Terminal-Bench results, matched against the coding agents in our index: accuracy, model pairing and cost, not vendor claims.
- [Buyer's guide](https://theagentsindex.com/blog/the-best-claude-code-alternatives-for-the-terminal) 27 Oct 2026 The best Claude Code alternatives for the terminal zot, Aider, OpenCode and GitHub Copilot CLI run Claude models from the terminal with a free plan, and add OpenAI models Claude Code lacks.
- [Buyer's guide](https://theagentsindex.com/blog/what-ai-agent-frameworks-cost-to-start-and-scale) 29 Oct 2026 What AI agent frameworks cost to start and scale Most AI agent frameworks are free to start, and the bill arrives with the hosted layer, priced by plan, seat, credits or usage.

## Is this yours?

Claiming is free. You get this page's reader numbers and the verified owner mark. You confirm it with an email at harvey.ai.

[Claim this listing](https://theagentsindex.com/claim/harvey-ai?from=owner-line) [Get Featured](https://theagentsindex.com/claim/harvey-ai?from=owner-line&next=featured)

[Report an error](mailto:hello@theagentsindex.com?subject=Report%20an%20error%3A%20Harvey%20(harvey-ai)&body=What%20is%20wrong%20on%20the%20page%2C%20and%20where%20on%20the%20product) if a fact is wrong.
