# GC AI — Legal AI platform for in-house counsel, spanning drafting, contract review, playbooks and research.

> Source: The Agents Index — https://theagentsindex.com/gc-ai (structured, researched, re-verified)
> Facts last verified: 2026-09-26

GC AI is a legal AI platform built for in-house legal teams, covering drafting, contract review, playbooks and case-law research. It is sold to corporate legal departments rather than law firms, and was founded by a three-time general counsel.

| Fact | Value |
| --- | --- |
| Website | https://gc.ai |
| Price | from $500/seat per month billed monthly, $416.67/seat per month billed annually |
| Free trial | Partly |

Self-serve 14-day trial at app.gc.ai/auth/sign-up, no sales call required. No standing $0 plan is published; after the trial the cheapest paid entry is the $500/month Individual seat. The REST API is separately described in the docs as 'free to test'.

## What it does

**Done-for-you service** · No — You buy and run the software yourself, self-serve from $500 a month; the vendor sells optional professional services for playbook buildouts and integrations, but does not staff or operate the legal work for you.

> Professional services are also available for playbook buildouts, custom integrations, workflows and support.
> — https://gc.ai/pricing · Common questions FAQ, 'Do you support enterprise features?' answer

## How much it does unattended

**Runs autonomously** · Partly — Automations run saved instructions on a schedule with no human present and cannot ask a clarifying question mid-run, but connector actions that send or change data default to needing approval, and the vendor frames output as work a lawyer still reviews.

> Automations are recurring GC AI chats. You set the task up once with the instructions, skills, files, playbooks, and context it needs, then pick a schedule.
> — https://docs.gc.ai/docs/automations/overview · Automations overview, opening paragraph

**Multi-agent** · Yes — Chat 2.0 runs specialised research and analysis processes in parallel and reconciles their findings before answering, so coordination happens inside one conversation rather than as user-assembled agents.

> When a complex legal question is asked, specialized research and analysis processes run in parallel. Findings are reconciled before output is delivered.
> — https://gc.ai/blog/introducing-chat-2-0-powered-by-a-multi-agent-architecture · Section 'First, Let's Talk About Agents'

**Agent permissions** · Yes — Permissions are set per action type rather than per app, so searching an inbox and sending mail are separate grants, and an in-chat card offers Approve, Always approve or Deny before anything leaves.

> Every action type has its own setting: Needs approval, Always allow, or Block. Searching your inbox and sending an email are different permissions, so GC AI reads freely
> — https://gc.ai/agent-connectors · 'You approve before anything happens' section, 'Set permissions action by action'

## Which models it runs on

**Claude** · Yes — Anthropic's Claude family is one of four providers the platform routes to, and Teams and Enterprise admins can additionally opt in to the Claude Fable 5 model, which carries 30-day retention rather than zero data retention.

> GC AI integrates multiple models from Anthropic's Claude family, ranging from highly capable reasoning models for complex legal analysis to faster models for lighter tasks.
> — https://docs.gc.ai/docs/chat/ai-models · AI Providers section, Anthropic subsection

**GPT** · Yes — OpenAI models are in the routing lineup for reasoning and analysis, and the vendor states a zero data retention agreement is in place with OpenAI.

> GC AI uses OpenAI's most advanced models for complex reasoning and analysis.
> — https://docs.gc.ai/docs/chat/ai-models · AI Providers section, OpenAI subsection

**Open models** · No — The provider inventory GC AI publishes covers only proprietary frontier vendors - Anthropic, OpenAI, Google and xAI, with Cohere for embeddings - and no open-weight model or customer-supplied checkpoint appears on any plan.

> In Settings, admins can switch a toggle to prevent use of one or more models from the following providers: Anthropic (Claude), OpenAI (GPT), and xAI (Grok).
> — https://docs.gc.ai/docs/chat/ai-model-opt-in-out · AI Model Opt Out section, opening paragraph (Google noted separately as not toggleable)

**Model choice** · Partly — The vendor picks the model per task and users cannot choose one per chat; the only customer control is a Teams or Enterprise admin switching whole providers off, and at least one provider must stay on.

> GC AI automatically selects the best model for each task. You don't need to choose a model manually.
> — https://docs.gc.ai/docs/chat/ai-models · Opening section, second paragraph

## Where you use it

**In your editor** · Partly — There is no native editor extension; an IDE reaches GC AI only by pointing Cursor's MCP configuration at the public-beta MCP endpoint, which an org admin must first enable.

> Add the server to your MCP configuration with the same URL. Cursor opens the GC AI sign-in on first use, then the tools are available to the agent.
> — https://docs.gc.ai/api-reference/mcp/connecting · 'Add the server to your host' section, Cursor tab

**On the command line** · Yes — A published npm CLI covers the whole REST surface from a terminal, authenticating with an API key, so scripts and shell pipelines are a first-class path.

> Every command maps to a REST endpoint, so anything you can do over the API you can do from your terminal, with tab-completable commands, typed flags, and human-readable output.
> — https://docs.gc.ai/api-reference/cli/overview · CLI Introduction, opening paragraph

**In a browser** · Yes — The hosted web app at app.gc.ai is the primary surface, carrying chat, files, projects and playbook reviews alongside browser-based document editing.

> plus interactive, day-to-day work across Slack, Microsoft Word, and the web app: chat, files, projects, playbook reviews, and the Slack connector, currently free with the seat
> — https://gc.ai/pricing · Common questions FAQ, 'What's included when I buy GC AI?' answer

## Whose machine it runs on

**Self-hosted** · No — Hosted SaaS only: buyers are told to allowlist the vendor's own domains for the web app and Word add-in, and GC AI states it runs all servers itself in the United States. No on-premise or customer-cloud deployment is offered on any plan.

> To ensure full functionality of GC AI products (Web App and Word Add-in), your network firewall or proxy must allow outbound connections to the following domains and ports.
> — https://docs.gc.ai/security/whitelisting · Whitelisting Requirements, opening paragraph above the Required Domains table

_Not established: Open source_

## What it costs to run

**How it meters** · Yes — Two meters run side by side: a per-seat subscription with unlimited chats and skills, plus pooled credits for API and MCP work priced per operation, so programmatic volume is billed separately from headcount.

> Usage is billed in credits, pooled across your organization. One credit is $5, which buys minutes of a $200/hr senior attorney's time. Each API chat message costs 1 credit.
> — https://gc.ai/blog/gc-ai-api · Section 'Credits, privacy, and control'

**Free tier** · Partly — You can start without talking to sales, but only as a 14-day trial; there is no standing zero-cost plan, and the cheapest permanent option is the $500 monthly seat.

> No, you can start a free, 14-day trial immediately or take a class.
> — https://gc.ai/pricing · Common questions FAQ, "Do I have to talk to sales? I'm an introvert." answer

**API access** · Yes — A public REST API covering chat, files, playbooks, projects and skills is open to any organisation once an admin enables it, with no extra seat, though usage draws pooled credits rather than being bundled into the subscription.

> GC AI's API is available to your organization, with access enabled by an organization admin. There's no extra seat required. Usage is billed in credits
> — https://gc.ai/pricing · Common questions FAQ, 'Does my plan include GC AI's API?' answer

**MCP server** · Yes — GC AI runs its own remote MCP server exposing twenty tools over OAuth, so another agent can query and act in the account; it is public beta and off until an org admin turns it on.

> The GC AI MCP server lets an AI assistant work directly in your GC AI account. Connect it once in a host like ChatGPT, Claude, or Cursor, and the assistant can search your files, run playbooks
> — https://docs.gc.ai/api-reference/mcp/overview · MCP Introduction, opening paragraph

_Not established: Bring your own key_

## Buying it for a team

**A company can buy it** · Yes — Team and Enterprise contracts exist with invoicing, procurement support and account executives, while a lone practitioner can still buy a single seat by card.

> Yes, invoicing options are available for team and enterprise plans. Single seats pay by card through our payment processor.
> — https://gc.ai/pricing · Common questions FAQ, 'Can I be invoiced?' answer

**Seat model** · Yes — Pricing is per seat and starts at one: $500 a month or $5,000 a year for the Individual plan. No seat minimum or ceiling is published, and the Team per-seat rate is quoted on request.

> GC AI publishes transparent pricing starting at $500/month for an individual seat, with monthly and annual subscriptions.
> — https://gc.ai/ · FAQ, 'How much does GC AI cost?' answer

**Pooled budget** · Partly — Seats are individual and carry unlimited chats, so nothing is shared there; only API and MCP consumption draws on one organisation-wide credit balance that admins monitor.

> Usage is billed in credits, pooled across your organization, so you only pay for what you use.
> — https://gc.ai/pricing · Common questions FAQ, 'Does my plan include GC AI's API?' answer

**Admin controls** · Yes — Admins gate what the agent may touch org-wide and per member - connectors disabled, read-only or full access, public chat sharing, Google Drive import, personal API keys and which model providers are used.

> For most connectors, organization admins control whether access is Disabled, Read-only, or Full access for all members or specific members in Settings → Policies → Connector policies.
> — https://docs.gc.ai/docs/organizations/permissions · Agent Connector access section

**Audit log** · Partly — Authentication audit logs and log streams are named as Enterprise features, and provider toggles plus connector connections record who changed what, but the docs still list advanced audit logging as not yet shipped.

> Yes. GC AI has SSO and MFA on Enterprise plans. Enterprise also includes Directory Sync, Domain Verification, Authentication Audit Logs, and Log Streams.
> — https://gc.ai/ · FAQ, 'Does GC AI Support SSO and MFA?' answer

**Single sign-on** · Partly — Both SAML and OIDC are supported through WorkOS with any major IdP, but SSO sits on Team and Enterprise only, needs domain verification first, and Individual seats get multi-factor authentication instead.

> GC AI supports SSO through WorkOS, which is compatible with any IdP that supports SAML or OIDC protocols.
> — https://docs.gc.ai/security/sso-setup · 'What is SSO?' section

## What happens to your code

**Opt out of training** · Yes — No opt-out is needed because the contractual default is no training on customer data, backed by zero data retention agreements; the exception is the optional Claude Fable 5 model, which retains data for 30 days for safety work.

> Our AI model providers are prohibited from using your data to train their models, and we maintain zero-data-retention agreements with our LLM providers wherever feasible.
> — https://gc.ai/security · FAQ, 'What data security does GC AI offer?' answer, Vendor Protection bullet

**Data residency** · Yes — Jurisdiction is stated plainly and is singular: every server sits in the United States, with standard contractual clauses covering cross-border transfers. No EU or other regional hosting option is offered.

> US-based infrastructure with all servers located exclusively in the United States
> — https://docs.gc.ai/security/overview · Enterprise-Grade Security bullet list, final bullet

**Getting out** · Partly — Chats export one at a time as PDF, DOCX or Markdown and tables as CSV or XLSX, with bulk download explicitly unavailable and full erasure by written request; unused credits are forfeited at the end of each term.

> Unused Credits expire at the end of each Subscription Term and do not carry forward to any renewal or subsequent term. Expired Credits have no cash value and are non-refundable.
> — https://gc.ai/terms · Exhibit A, Credit Plans, clause 1.4 Expiration

**Certifications** · Yes — SOC 2 Type II, SOC 3 and GDPR are all claimed, with the SOC 2 report available under NDA through the Trust Center and a DPA included in the standard terms for every customer.

> Yes. GC AI is SOC 2 Type II and SOC 3 certified, GDPR compliant, with zero data retention agreements with OpenAI and Anthropic, and AES-256 encryption.
> — https://gc.ai/ · FAQ, 'Does GC AI meet security standards?' answer

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

Pages about this product that the vendor does not control. The label states how independent each one is.

| Source | Kind | Read |
| --- | --- | --- |
| artificiallawyer.com | Independent source | 2026-08-23 |
| vaquill.ai | Independent source | 2026-08-23 |

## What changed

- 2026-09-26 — Free tier
- 2026-09-14 — Capabilities · Commercial terms · Free tier · Price
- 2026-09-02 — Capabilities · Commercial terms · Free tier
- 2026-08-31 — Capabilities · Commercial terms · Free tier · Price
- 2026-08-22 — Alternatives · LinkedIn link · One-liner · Summary

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent GC AI, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.