Photon’s iMessage API, and where it fits

What Photon is

Photon is a messaging platform aimed at AI agents, built around Spectrum, an open-source TypeScript SDK. The pitch is one agent model deployed across several messaging surfaces — iMessage first, plus WhatsApp, Telegram, and SMS/RCS fallback — rather than an integration per channel.

Its iMessage platform page describes managed lines and enterprise-grade infrastructure, and the company cites production deployments and a high daily API call volume. Treat vendor-published scale claims as positioning rather than as verified benchmarks, but the product is substantial and the developer surface is real.

The feature surface is unusually deep

Photon publishes a longer list of native iMessage behaviors than most of this market: styled text with bold, italic, rich links, and text and bubble effects; photos, Live Photos, video, documents, and multi-image carousels; native reactions and threaded replies; typing indicators; voice note bubbles; permissioned real-time location sharing; native contact cards; group chats with custom names and photos; native polls; and custom conversation backgrounds.

It also describes Mini Apps and App Clips for interactive UI inside a message, voice calling on the same number as messaging, and SIP integration. If your product depends on the richer end of what iMessage can express, this is the most expansive published feature list in the category.

• Styled text, text effects, and bubble effects
• Live Photos, video, documents, and image carousels
• Native reactions, threaded replies, and typing indicators
• Group chats with custom names and photos, and native polls
• Mini Apps and App Clips for in-message interactive UI
• Voice calling on the messaging number, plus SIP

What Photon publishes for pricing

Photon publishes four tiers. These figures come from its own pricing page, retrieved on August 27, 2026 — confirm current terms before committing.

The important structural detail is that the cheap tiers are priced per user and the production tier is priced per line. Those are different units, and a comparison that puts $25 beside another vendor’s dedicated-line price is not comparing like with like.

• Free — $0, up to 10 users, unlimited daily messages, direct messaging API, SMS/RCS and Telegram included, community support
• Pro — $25/month, up to 100 users, same messaging surface, community support with fast-track ticketing
• Business — $250/line/month, dedicated lines, group messaging API, cold outreach capped at 50 new contacts per day, WhatsApp and phone call channels, dedicated Slack and Discord support
• Enterprise — custom, unlimited messages with the lowest throttling, dedicated lines, priority support
• No setup fee is published on any tier

Read the cold-outreach limit carefully

The Business tier publishes 50 new contacts per day per line for cold outreach. That is a meaningful production ceiling and one of the clearer disclosures in this market — several competitors leave the equivalent number to a sales conversation.

It also sets the shape of your bill. A team needing 200 new conversations a day needs four lines, which is $1,000 a month before any other channel. Model your peak day against that number before comparing headline prices.

The architecture is genuinely different

Photon describes its channel as a single persistent gRPC stream between its cloud, which owns the iMessage line, and a small Node sidecar running next to your agent — no public URL, no webhook endpoint, no Mac relay on your side.

That is a real advantage for an agent deployment: no inbound endpoint to expose, verify, and keep available. It is a real constraint everywhere else — a persistent stream and a sidecar process is a heavier operational commitment than a webhook, and it assumes your architecture has somewhere to run a long-lived Node process. Serverless-first stacks should check that assumption early.

• No public webhook URL to expose or verify
• A long-lived sidecar process next to your agent
• Persistent stream rather than request-response delivery
• Well suited to a running agent, awkward for a cron job or a serverless function

Where Photon fits best

Choose Photon when the deliverable is an agent and iMessage is one of several places it needs to appear. If you genuinely need the same agent on iMessage, WhatsApp, and Telegram, building that once against one SDK is a substantial saving over three integrations and three sets of delivery semantics.

It is also the strongest published option if your product depends on the richer iMessage surface — polls, Mini Apps, in-message interactive UI, voice on the same number. Nobody else in this market publishes that list.

Where a dedicated line fits better

Choose a conventional managed line when the deliverable is a business phone number rather than an agent: a sales team following up, a service business coordinating appointments, a support line several teammates work together.

The decisive question is usually whether people need to be in the conversation. An agent SDK optimizes for code owning the thread; a Message Center optimizes for a person taking it over mid-conversation, seeing what the automation said, and answering as the business. Both matter in a real operation, but products weight them differently.

• Do teammates need to read and take over live threads?
• Is the primary artifact a phone number or an agent?
• Do you need one channel done well, or several channels at once?
• Is there somewhere to run a persistent process?
• Is outbound volume above 50 new contacts per line per day?

How the published prices line up

Here is what providers published as of late August 2026, from their own pricing pages. Units differ, so compare dedicated line to dedicated line.

Photon’s $250 Business line sits at the higher end of published dedicated pricing, which is coherent given the multi-channel surface — you are buying more than an iMessage line. Whether that is good value depends entirely on whether you use the other channels.

• Miss Blue — $39/month shared testing number; $99/month per dedicated number with unlimited commercial usage, contacts, team access, Message Center, and API access
• Photon — Free; $25/month Pro up to 100 users; $250/line/month Business; Enterprise custom
• Sendblue — free sandbox; $100/month per line inbound-first; outbound on Enterprise quote
• Blooio — $39 Starter, $89 Commercial Shared, $109 Inbound, $289 Commercial Dedicated
• Project Blue — $300/month plus $500 setup, or $166.50/month billed annually with setup waived
• Blue Reacher — $249/month per line plus $449 one-time activation, waived on annual
• Tuco AI — from $149/month plus a $335 one-time setup fee
• Claw Messenger — from $5/month on a metered message allowance
• Linq Blue — no published production figure

Questions worth asking

The published documentation answers more than most, which narrows the remaining questions to operations and exit. Ask these of Photon and of everyone else on your list.

Pay particular attention to throttling. The Enterprise tier is described as having the lowest throttling, which implies the lower tiers are throttled — worth understanding precisely before you size a deployment on the Business plan.

• What throttling applies below Enterprise, and at what thresholds?
• Does the 50-new-contacts-per-day limit apply per line or per account?
• What do WhatsApp and phone-call channels cost on Business?
• Who owns the number if we leave, and can it be ported?
• What happens to the agent if the sidecar disconnects — buffered or dropped?
• Is there a human inbox, or is the thread code-owned by design?
• What is the support commitment during a delivery incident?

Where Miss Blue sits

Miss Blue is the other shape: a dedicated blue line, an HTTP API with durable conversation events and signed webhooks, and a shared Message Center where teammates read the same thread and take over from an automation. Published pricing is $39 per month for a shared testing number and $99 per month per dedicated number, with unlimited commercial usage, contacts, team access, the Message Center, and API access included on the dedicated plan.

The displayed $39 and $99 figures are current 50%-off sale rates against regular prices of $78 and $198, locked in while the plan remains active. If you are building one agent for four channels, Photon is the more natural fit. If you need a business line that a team works and an API your product calls, compare on that basis — and if you want the side-by-side, the comparison page covers it.