GC AI

What it does

Done-for-you service

PartlySoftware you run, with optional paid professional services for playbook buildouts and integrations, plus Solutions Attorney support on Team.Professional services are also available for playbook buildouts, custom integrations, workflows and support.gc.ai/pricing ↗

How much it does unattended

Runs autonomously

PartlyAgentic loops run up to 50 steps per turn and Automations run on a schedule unattended, but human gates and per-action approvals hold back anything that sends or changes data.GC AI autonomously decides to take multiple steps to accomplish a task, rather than giving a one-shot answer.gc.ai/blog/how-gc-ai-thinks-the-agentic… ↗

Multi-agent

YesChat 2.0 runs an orchestrator node that fans work out to parallel research and playbook-check agents and reconciles their findings before answering.In Chat 2.0, agents operate at the architectural layer within the core chat experience. They coordinate research and analysis behind the scenesgc.ai/blog/introducing-chat-2-0-powered… ↗

Agent permissions

YesPer-action setting on every connector tool, defaulting to Needs approval for anything that sends or changes data; in-chat cards offer Approve, Always approve or Deny.Every action type has its own setting: Needs approval, Always allow, or Block.gc.ai/agent-connectors ↗

Which models it runs on

Claude

YesMultiple Anthropic Claude models are in the routing lineup; admins on Teams and Enterprise can opt in to Claude Fable 5.GC AI integrates multiple models from Anthropic's Claude family, ranging from highly capable reasoning models for complex legal analysis to faster…docs.gc.ai/docs/chat/ai-models ↗

GPT

YesOpenAI models are in the lineup, with zero-data-retention terms named alongside Anthropic.GC AI uses OpenAI's most advanced models for complex reasoning and analysis.docs.gc.ai/docs/chat/ai-models ↗

Open models

NoThe provider list is closed and entirely proprietary: Anthropic, OpenAI, Google and xAI for chat, Cohere for embeddings only. No open-weight model is named anywhere.GC AI uses advanced large language models (LLMs) from leading AI providers to deliver features like contract analysis, legal research, and policy…docs.gc.ai/docs/chat/ai-models ↗

Model choice

PartlyThe vendor routes: no per-chat model picker, with cross-provider fallback chains. Admins on Teams and Enterprise can opt providers in or out, which is a veto rather than a choice.GC AI automatically selects the best model for each task. You don't need to choose a model manually.docs.gc.ai/docs/chat/ai-models ↗

Where you use it

In your editor

PartlyNo native IDE extension. It is reachable from Cursor and other MCP hosts through the MCP server (public beta), and its document-editor surface is the Microsoft Word add-in.Connect it once in a host like ChatGPT, Claude, or Cursor, and the assistant can search your files, run playbooks, ask legal questions grounded in…docs.gc.ai/api-reference/mcp/overview ↗

On the command line

Yesgcai, an npm-installed command-line client covering every REST endpoint, with --json output for piping and scripted use.gcai is the command-line client for the GC AI API. Every command maps to a REST endpoint, so anything you can do over the API you can do from your…docs.gc.ai/api-reference/cli/overview ↗

In a browser

YesHosted web app at app.gc.ai for chat, files, projects and playbooks, plus Easy Edit for browser-based document editing.Your seat includes market-leading core legal AI, web research, a Word add-in for commercial contract review, and Easy Edit for browser-based document…gc.ai/pricing ↗

Whose machine it runs on

Self-hosted

NoVendor-hosted SaaS on GC AI's own US servers; customer network work is allowlisting app.gc.ai and word.getgc.ai, not deploying anything.US-based infrastructure with all servers located exclusively in the United Statesdocs.gc.ai/security/overview ↗

Not established: Open source

What it costs to run

How it meters

YesTwo meters, both published: a per-seat subscription from $500/month, and credits at $5 each for API and MCP work, pooled across the organisation.Each API chat message costs 1 credit. Each playbook check costs 0.1 credit, so a 30-check playbook review costs 3 credits.gc.ai/blog/gc-ai-api ↗

Free tier

PartlyA 14-day free trial, self-serve and without a sales call, and the API is free to test, but there is no permanent free tier.Yes, you can start a 14-day free trial here.gc.ai/pricing ↗

API access

PartlyA REST API for any organisation, but an admin must enable it and usage is billed in credits rather than bundled into the seat; no extra seat required.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…gc.ai/pricing ↗

MCP server

YesAn MCP server exposing 20 tools to ChatGPT, Claude, Cursor and other hosts, scoped to the signed-in user; OAuth client-credentials can be minted for machine clients. Documented as public beta.The GC AI MCP server lets an AI assistant work directly in your GC AI account.docs.gc.ai/api-reference/mcp/overview ↗

Not established: Bring your own key

Buying it for a team

A company can buy it

YesTeam and Enterprise plans are sold to organisations with invoicing and procurement support; a single seat self-serves by card.Yes, invoicing options are available for team and enterprise plans. Single seats pay by card through our payment processor.gc.ai/pricing ↗

Seat model

YesPer seat from $500/month with a one-seat floor and no published ceiling; Team and Enterprise seat pricing is quoted, and seats can be added later by upgrading.GC AI publishes transparent pricing starting at $500/month for an individual seat, with monthly and annual subscriptions.gc.ai ↗

Pooled budget

PartlyAPI and MCP credits draw from one organisation-wide balance that admins monitor; the subscription itself is priced per seat, so seats do not share anything.Usage is billed in credits, pooled across your organization, so you only pay for what you use.gc.ai/pricing ↗

Admin controls

YesAdmins gate Agent Connectors org-wide, see who connected which apps, opt AI providers in or out, manage web-search approval and disable public chat sharing.Yes. Admins can enable or disable Agent Connectors org-wide under Settings → Integrations.gc.ai/security ↗

Audit log

PartlyAuthentication audit logs and log streams are Enterprise-only. Product-wide there is narrower recording: model opt-in/out changes are logged with who and when, and connector policies show connections only, not content.Enterprise also includes Directory Sync, Domain Verification, Authentication Audit Logs, and Log Streams.gc.ai ↗

Single sign-on

PartlySAML or OIDC against any major IdP via WorkOS, requiring domain verification, but gated to the paid team tiers. The Individual plan gets MFA only.GC AI supports SSO through WorkOS, which is compatible with any IdP that supports SAML or OIDC protocols.docs.gc.ai/security/sso-setup ↗

What happens to your code

Opt out of training

YesNo training is the default rather than a setting: providers are contractually barred from training on customer data, with zero-data-retention agreements named for OpenAI and Anthropic.Our AI model providers are prohibited from using your data to train their models, and we maintain zero-data-retention agreements with our LLM…gc.ai/security ↗

Data residency

YesUnited States only, with no EU or other region offered. Cross-border transfers run under Standard Contractual Clauses, and the platform is used in jurisdictions worldwide from that US footprint.GC AI and all our servers are located exclusively in the United States.docs.gc.ai/security/data-protection ↗

Getting out

PartlyChats, individual messages and tables export to PDF, DOCX, Markdown, CSV or XLSX and playbooks to DOCX, but only one at a time, with no bulk export. Deletion is self-serve, org-wide erasure is by email. Whether unused credits expire is not published.Chats must be downloaded individually and cannot be bulk downloaded.docs.gc.ai/docs/chat/how-to-download ↗

Certifications

YesSOC 2 Type II and SOC 3 certified and GDPR compliant, with reports in a public trust center and the SOC 2 report available under NDA.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…gc.ai ↗