Linq is a messaging API that lets an app or AI agent text people over native iMessage, RCS and SMS from a single integration. Several consumer texting assistants run on it, including Caddy and Pally, and reports say OpenAI's dots texts route through it as well. If you have texted an AI assistant from an iPhone and seen blue bubbles, there is a fair chance an API like Linq was in the middle.
Most of what follows comes from Linq's own website and blog, plus press coverage of its funding. Figures marked as company figures have not been independently verified.
What does Linq do?
Linq describes itself as a conversational messaging API for iMessage, RCS, SMS and voice. A developer connects an agent to Linq, gets a phone number, and the agent can send and receive messages. When the recipient has an iPhone, the message can arrive as iMessage. On Android it can use RCS, with SMS as the fallback.
According to Linq's website, the API supports:
- Native iMessage and RCS, with SMS fallback
- Reactions, voice notes and rich media
- Typing indicators
- Group chats
- A command-line tool (
npm install -g @linqapp/cli) and a sandbox for testing
- AI agent
- Linq API
- iMessage if iPhone
- RCS if supported
- SMS fallback
For an AI assistant, those features matter. Typing indicators make a slow model reply feel less broken, voice notes let people talk instead of type, and group chat support enables the kind of AI agent group chats that plain SMS handles badly.
What does Linq claim about scale and reliability?
- Series A, reported by TechCrunch in Feb 2026
- $20M
- Monthly active users across customers' agents (company figure)
- 134K
- Messages a month (company figure)
- 30M+
- Uptime Linq says is backed by a contractual SLA
- 99.95%
Linq also says it has sub-120ms latency, more than 1,000 teams as customers, and SOC 2 Type II certification, which it describes as making it the only SOC 2 Type II iMessage API. That last claim is Linq's own and we have not checked it against competitors. Its blog page displays a figure of more than 900M messages, which presumably counts total messages over time, though the page does not say.
On pricing, Linq's site lists no prices as of November 2026. Its marketing says it does not use per-message pricing and offers a seven-day free trial.
Which texting agents use Linq?
Linq's blog post on Caddy moving its assistant to texting, dated September 4, 2026, is the most detailed public account. Caddy's founders started with a voice-driven Mac app, ran the same backend over email, Slack and iMessage, and found iMessage took off. After moving to Linq, one founder said usage "just spiked". Caddy did not share retention or volume numbers.
The same post cites two other customers:
- VentNow ran tests and found users in native texting sent about 20% more messages per session and were about 40% more active daily than in its app.
- Lindy's founder said "the overwhelming majority of all traffic happens over iMessage."
These are vendor-published customer stories, so read them as directional. Our startup texting agents comparison covers several of these assistants from the user's side.
Does OpenAI's dots use Linq?
Reports say so. Windows Report and gagadget, citing OpenAI's support documentation, say dot texts are sent through Linq and are not end-to-end encrypted, even over iMessage or RCS. OpenAI's own channels page does not name Linq or discuss encryption. Dots texting has been rolling out in beta to US Pro users since October 28, 2026. See our guide to texting OpenAI dots and getpersonalaiagents.com's dots review.
How is Linq different from Apple Messages for Business?
They are two different ways for a business to appear in the Messages app.
| Linq-style iMessage API | Apple Messages for Business | |
|---|---|---|
| Who runs it | Third-party provider | Apple |
| How chats start | Texting a phone number | Started by the customer in Messages |
| Approval | Provider onboarding | Apple verifies the business, which connects through an approved messaging provider |
| AI agents | Used by several startup assistants | Poke was the only approved AI agent as of June 2026 |
| Works on Android | Yes, through RCS or SMS fallback | No, Apple devices only (iPhone, iPad, Mac, Apple Watch) |
Our Apple Messages for Business guide explains Apple's route in detail, and best AI agent for iMessage compares the assistants available today.
What should you check before building on an iMessage API?
Pros
- Native iMessage features such as reactions, typing indicators and voice notes
- One integration covering iMessage, RCS and SMS
- Group chat support
- Vendor-published stories suggest higher engagement than in-app chat
Cons
- No published prices
- Encryption claims are about the provider, not end-to-end
- Scale and security claims are self-reported
- Depends on a third party's access to iMessage
A few questions worth asking any provider in this category:
- Encryption. Who can read message content in transit and at rest? "Encrypted and inaccessible to us" is a provider claim, and not the same as end-to-end encryption. Our encryption guide for texting agents covers what each channel protects.
- US compliance. If any traffic falls back to SMS from a US number, ask how the provider handles A2P registration. See our 10DLC registration guide.
- Data retention. How long are messages stored, and can users delete them? Our guide to deleting texting agent data covers what users can ask for.
- Continuity. iMessage access through a third party depends on that provider's arrangements. Ask what happens to your numbers and history if service changes.
Bottom line
Linq is an iMessage, RCS and SMS API used by several AI texting assistants, and reportedly by OpenAI for dots. It raised a $20M Series A reported in February 2026, and its own figures put its customers at 134K monthly active users and more than 30M messages a month. It offers native iMessage features that plain SMS cannot, but it publishes no prices, and its security and scale figures are self-reported. If you are building a consumer texting agent, it is one of the main routes onto the iPhone. Apple Messages for Business is the other, with tighter approval.


