Last updated: 2026-10-04
Cassidy is a no-code platform for building AI agents and multi-step workflows grounded in a company's own documents, wikis and CRM records. It offers 10 deployment options, from Slack and Microsoft Teams to Word, Excel and an API, and lets each workflow step use models from OpenAI, Anthropic or Google.
About Cassidy
Cassidy is a no-code AI automation platform founded in 2023 by Justin Fineberg and Ian Woodfill in New York. It raised a $3.7M seed round in 2024 led by The General Partnership, then a $10M Series A announced in September 2025 and led by HOF Capital, for $13.7M in total.
The product is built around a knowledge base. Teams connect Google Drive, SharePoint, Box, Notion, Confluence and CRM data from HubSpot or Salesforce, and Cassidy turns those files into text pages that its agents read before they answer. Agents are chat assistants that cite that context. Workflows are trigger-based automations: a scheduled run, a CRM update or a new ticket in Zendesk starts a chain of AI steps, conditional branches, retries and human sign-offs that ends in an action in another app.
Cassidy does not tie a workflow to one model. Each step can call models from OpenAI, the maker of ChatGPT, from Anthropic, which builds Claude, or from Google, so a team can use one model for drafting and another for analysis. Cassidy says it keeps customer files out of model training, and its privacy notice says data is stored on servers in the United States.
The buyers Cassidy courts on its own site are insurance brokers and carriers, employee-benefits firms, industrial companies and professional-services teams, with proposal and RFP work, sales enablement and support triage as the recurring jobs. That focus shows in features such as an answer hub for pre-approved RFP responses, OCR parsing for scanned documents and a document-verification step that keeps outdated files out of a workflow.
Where the alternatives win: Zapier and Make connect to far more apps and publish self-serve paid plans, so a small team can buy more capacity with a card instead of a sales call. n8n can be self-hosted, which suits developer-led teams that want to own the runtime. Glean is the closer fit when the main job is searching company knowledge rather than automating work, and Microsoft Copilot already sits inside Microsoft 365 for teams that only need drafting help. Cassidy is also the wrong pick for a team without organised documents, because its answers are only as good as the knowledge base behind them. More options sit in the AI agent builders category.
Pricing
Cassidy lists two plans on its pricing page, checked 4 October 2026. Starter is free and Business is custom-priced after a demo, with seats, AI credits and storage sized to the team. No self-serve paid tier is published, so the $49 Pro plan that older listings quote no longer exists.
Credits can be topped up at any time, and Cassidy says it alerts admins near the limit rather than stopping a workflow. One inconsistency on that page: the Starter plan card lists 30,000 knowledge pages while the comparison table says up to 100,000, so we quote the plan card. For self-serve paid alternatives, see Zapier or Make.
| Tier | Monthly price | What it includes |
|---|---|---|
| Starter | Free | 3 seats and 1 workspace, 10,000 AI credits a month, 5 agents and 5 workflows, 30,000 knowledge pages, Source sync every 24 hours |
| Business | Custom | Custom seats, workspaces, credits and storage, Instant source sync, Access to all features, Premium onboarding and support |
Feature Comparison by Tier
| Feature | Starter | Business |
|---|---|---|
| Seats | 3 | Custom |
| Workspaces | 1 | Custom |
| AI credits per month | 10,000 | Custom |
| Agents | 5 | Custom |
| Workflows | 5 | Custom |
| Knowledge pages | 30,000 | Custom |
| Source sync | Every 24 hours | Instant |
| Model choice | Limited | Full |
Key Features
- Context-Grounded Agents: Agents answer from your own documents, wikis and CRM records and cite the source page, with each 20-page PDF counting as 20 knowledge pages.
- Multi-Step Workflows: Trigger-based automations combine scheduled or event triggers, conditional branches, retries and actions in connected apps, with version history to roll back a change.
- Human-in-the-Loop Steps: Approval points can sit anywhere in a workflow, so a person signs off on a proposal or reply before the next action runs.
- Bulk Execution: One workflow can run across a large dataset, which Cassidy illustrates as processing 1,000 documents at once.
- Knowledge Base Sync: Google Drive, SharePoint and other connected sources re-sync every 24 hours on Starter and instantly on Business.
- Deploy Where Teams Already Work: Agents run inside Slack, Microsoft Teams, Google Chat, Gmail, Outlook, Word, Excel, a Chrome or Edge extension, a browser embed and the API.
Pros
- AI reasoning sits inside every workflow step rather than as an add-on, so an automation can read company context before it acts instead of only moving data between apps.
- The compliance set (SOC 2 Type II, GDPR, HIPAA, CASA) makes Cassidy a realistic option in regulated fields such as insurance and benefits, where lighter automation tools often fail procurement review.
- Agents meet staff where they already work, from Slack and Teams to Word, Excel and Outlook, so adoption does not depend on people opening a new tab.
Cons
- There is no self-serve paid plan: once a team outgrows Starter, the next step is a demo and a custom quote, which slows small teams that just want more credits.
- Starter re-syncs connected sources only once a day, so an agent can answer from yesterday's version of a document until the team moves to Business.
- Effective automation depends on the quality and structure of existing company data; teams without organised document repositories get limited benefit from the knowledge base features.
Data Handling
- Trains on your data
- No
- Training-data policy
- Cassidy states customer data is never used to train AI models (pricing and security pages, checked 2026-10-04).
- Data residency
- United States (privacy notice, updated 2026-03-17)
- Compliance
- SOC 2 Type II · GDPR · HIPAA · CASA Certified
Frequently Asked Questions
How much do you pay for Cassidy?
Nothing on Starter, the free plan; Business is custom-priced after a demo. Cassidy's pricing page, checked 4 October 2026, lists no self-serve paid tier, so the $49 Pro plan older listings quote is gone. Business sizes seats, AI credits and storage to the team and adds instant source sync and premium onboarding, and extra credits can be bought as top-ups on either plan.
Does Cassidy have a free plan?
Yes. Starter is free for 3 seats sharing one workspace, with a monthly allowance of 10,000 AI credits, room for 5 agents and 5 workflows, and 30,000 knowledge pages, per Cassidy's pricing page on 4 October 2026. Connected sources re-sync every 24 hours and model choice is limited on Starter. The same page's FAQ also mentions a 14-day free trial for evaluating the wider platform.
What are Cassidy's closest competitors?
The main alternatives are Zapier, which reaches far more third-party apps but does not ground each step in a company knowledge base; n8n, which is self-hostable and code-first for developer-led teams; and Make, a visual builder aimed at connecting apps. Microsoft Copilot is the pick for teams that only need drafting help inside Microsoft 365. Cassidy fits best when a team wants AI judgment embedded directly into the workflow.
What separates Cassidy from Zapier?
Zapier connects to far more third-party apps and relies on simple point-to-point triggers, but each step just moves data with no built-in AI reasoning. Cassidy grounds every step in the business's own knowledge base, so workflows can draft content, triage tickets, or answer questions using real company context instead of passing data untouched. Teams that mainly need broad app coverage should pick Zapier; teams that need AI judgment inside the workflow should pick Cassidy.
How long does it take to get going with Cassidy?
A first agent can be live within one sitting. Sign up for Starter at cassidyai.com, connect one data source such as Google Drive or Notion so the AI has company context, then build an agent or workflow from a template and test it before turning on triggers. Larger rollouts with SSO and live sync go through a demo for the Business plan.
Top Alternatives
- Zapier: Zapier carries a larger third-party app catalog; Cassidy carries AI reasoning grounded in a company's own knowledge base at every step.
- n8n: n8n stays self-hostable for teams that want to code custom logic; Cassidy stays no-code for teams that would rather non-technical staff run it themselves.
- Make: Make offers a pure visual workflow canvas for connecting apps; Cassidy instead reasons over the company's own documents and records at each step rather than just moving data between them.
HokAI guides covering Cassidy
- Best AI Agent Builders in 2026: Automate, Build or Self-Host: Re-checked 18 Oct 2026: Lindy, Cassidy and Copilot Studio automate work, Atoms and Emergent build apps, and three open-source names have gone stale for good.
- HubSpot AI Alternatives: What Agent Hub and Breeze Cover (and What They Don't): HubSpot's AI agents bill in credits: 50 for a resolved chat, 100 for a lead. What that buys on each plan, and nine alternatives by job, checked 19 Oct 2026.