Nango: The Integration Layer for AI Agents, Explained
Nango is a platform that lets software, and the AI agents inside it, connect to other apps on behalf of users. A product built on Nango lets its users connect accounts like Gmail, Salesforce or Linear, and Nango handles the hard parts: OAuth and API keys, token refreshes, storing credentials, running the integration code, retries, rate limits and logs. It supports more than 1,000 APIs and MCP servers, with its homepage now citing 1,200+, and its code is on GitHub with about 13,000 stars. Nango is used by companies including Replit, Mercor and Ramp.
The quick answer: Nango is infrastructure for developers building integrations into their own products and agents, not an app end users install. Its pitch is that you keep integrations as code you own, generated with AI coding tools, while Nango runs the auth and plumbing. It has a free tier, usage-based pricing, SOC 2 Type II and HIPAA support, and a self-hosting option. A disclosure: we make Carly, an AI assistant that connects to thousands of apps for its users, so we work on a related problem from the other side. Everything below comes from Nango’s site, docs and announcements.
Nango at a glance
Checked against Nango’s site, documentation, pricing and GitHub repository on November 12, 2026.
| Nango | |
|---|---|
| What it is | An integration platform for products and AI agents: auth, tools, syncs, webhooks and MCP |
| APIs supported | 1,000+ in the docs, 1,200+ APIs and MCPs on the homepage |
| Built for | Engineering teams building integrations into their own apps and agents |
| Code | On GitHub under the Elastic License 2.0, about 13,000 stars; self-hostable |
| Pricing | Free tier; pay-as-you-go from $50 a month; Enterprise custom |
| Compliance | SOC 2 Type II, GDPR, HIPAA with a BAA |
| Company | Co-founded by Robin Guldener and Bastien Beurier; Y Combinator W23; $7.5M seed led by Gradient |
What Nango does
Nango’s docs describe it as “the integration layer for AI-built, code-owned product integrations.” The idea is that every AI product needs to connect to the tools its users already work in, and building each integration from scratch, with its own auth, data syncs, rate limits and edge cases, used to take weeks.
Nango splits the work into two core pieces:
- Auth, for connecting user accounts. You embed Nango’s connect screen in your product, and it handles OAuth, API keys, token refresh, scopes, permissions and credential storage.
- Functions, for the integration logic. These are TypeScript functions that call an app’s API, transform data, sync records, process webhooks or run on demand. You generate them with an AI coding agent, then edit, test and version them like any other code.
Your app, background jobs or AI agents then use those functions through Nango’s API, SDKs or MCP server.
What you can build with it
Nango’s docs list the main uses:
- Tool calling for AI agents. Give agents scoped, observable access to external APIs.
- Syncing data. Keep data from other apps fresh in your product or a retrieval pipeline, with incremental syncs, checkpoints and deletion detection.
- On-demand actions. Read, write and automate other apps’ APIs when your product needs to.
- Webhooks. Receive webhooks from other apps and forward or process them.
- Unified APIs. Put several similar APIs behind one stable interface of your own.
- Per-customer customization. Adjust how an integration behaves for each customer without forking code.
Nango says it has more than 10,000 pre-built tools, triggers and syncs, which you can use as-is or extend.
Agent sessions
Nango’s newest feature is aimed squarely at AI agents. Agent sessions are “an optimized interface for AI agents to use integrations,” and each one has three parts: a tenant, meaning whose accounts the agent can act on; a toolset, meaning which integrations and tools it can use; and discovery, meaning how it finds and calls them.
Your backend creates the session and exposes it to the agent as an MCP server. Two rules matter for security: “The agent never sees raw credentials, and cannot change the scope of its own session.” Nango’s examples include a sales agent with one user’s Gmail and Google Calendar plus read-only access to the company’s Salesforce, and a developer agent with a team’s Linear and a user’s personal GitHub.
The homepage adds governance features around this: permissions and scopes, isolated sandboxes, logs and an audit trail, along with session memory that helps an agent recover from errors.
Pricing
Nango’s pricing is usage-based:
| Plan | Price | What it includes |
|---|---|---|
| Free | $0 | Auth, tools and syncs with hard-capped limits that reset monthly |
| Pay-as-you-go | $50 a month | $50 of credit each month, then usage rates |
| Enterprise | Custom | Commitment discounts, self-hosting, MSA and SLA |
The free tier includes 10 connections, 10 hours of compute and 10 GB of data transfer a month. Beyond that, a connection, meaning one user’s authorized account, costs $0.29 a month, compute costs $0.72 an hour, and data transfer costs $0.50 per GB. The pricing page also lists features such as removing Nango branding from the connect screen, role-based access, SAML SSO, SCIM, audit trails and HIPAA with a business associate agreement, some of which depend on the plan.
Open source and self-hosting
Nango’s code is public on GitHub, where the repository has about 13,000 stars and 1,500 forks. Nango calls itself open source and self-hostable. Strictly, the license is the Elastic License 2.0, a source-available license that lets you use, modify and self-host the software but not offer it to others as a hosted or managed service. For most teams running Nango inside their own product, that distinction doesn’t matter. For anyone planning to resell it as a service, it does.
Self-hosting and bring-your-own-cloud deployments are offered with support on the Enterprise plan.
The company and traction
Nango was co-founded by Robin Guldener and Bastien Beurier and went through Y Combinator’s Winter 2023 batch. In March 2026, Nango announced a $7.5 million seed round led by Gradient, with Horizon, Y Combinator and angel investors including Akshay Kothari of Notion and John Kim of Sendbird.
In that announcement, Nango said more than 10,000 engineering teams had used it, that it processes billions of API requests a month, and that it “already generates several million in ARR” and was cash-flow positive when it raised. It named Replit, Mercor and Ramp as customers, and said Motion had built hundreds of integrations with it in weeks. Those figures are Nango’s own. Its homepage quotes engineers at Replit, Vapi, Plain and Electric.
How Nango compares
Nango competes with other integration platforms for AI agents, including Composio, Arcade, Paragon and Pipedream Connect. Our Composio alternatives guide compares them. Nango’s distinguishing traits are that integrations live as code you own and review, its source is public and self-hostable, and it covers data syncs and webhooks as well as agent tool calls.
Nango is for teams building integrations into their own software. If you want an assistant that’s already connected to the apps you use, Carly connects to thousands of apps, labels and sorts email across all your Gmail and Outlook accounts by your rules, handles advanced scheduling, and makes and answers calls and texts, from inside ChatGPT or Claude or by text.
Who Nango suits
Nango is a good fit if:
- You’re building a product or AI agent that needs to connect to your users’ accounts in other apps.
- You want integrations as code you control, generated with AI coding tools.
- You need data syncs and webhooks, not just agent tool calls.
- You want the option to self-host, or need SOC 2 and HIPAA.
Look elsewhere if:
- You’re an end user looking for an assistant, not a developer building one.
- You want to offer integration infrastructure to others as a hosted service, which the Elastic License restricts.
- You’d rather use only pre-built connectors with no code at all.
FAQ
What is Nango?
Nango is an integration platform that handles authentication, tool calls, data syncs and webhooks for more than 1,000 APIs, so developers can connect their products and AI agents to the apps their users already use.
Is Nango open source?
Nango’s code is public on GitHub and self-hostable, under the Elastic License 2.0. That license allows using and modifying it but not offering it to others as a hosted service.
How much does Nango cost?
Nango has a free tier with 10 connections, 10 compute hours and 10 GB a month. Pay-as-you-go is $50 a month with $50 of credit, then $0.29 per connection, $0.72 per compute hour and $0.50 per GB. Enterprise is custom.
Does Nango work with MCP?
Yes. Nango can expose integrations to AI agents as an MCP server through agent sessions, and can let users connect to MCP servers through its auth flow.
Who uses Nango?
Nango names Replit, Mercor, Ramp and Motion among its customers, and says more than 10,000 engineering teams have used it.
Who founded Nango?
Robin Guldener and Bastien Beurier co-founded Nango. It’s a Y Combinator W23 company and raised a $7.5 million seed round led by Gradient in March 2026.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
