MorseSign inStart freeOpen Morse

Morse vs OpenRouter: a private chat app or a developer's model gateway

The Morse teamPublished Facts checked 11 min read

Morse vs OpenRouter is a choice between a private chat app and a developer's gateway to hundreds of models that happens to have a chat window. As of November 2026, OpenRouter proxies each prompt under your account to whichever provider its router picks for that request, and the privacy you get depends on the settings you've chosen. Morse strips your identity from GPT, Claude, and Gemini, runs open models on hosts that agreed to keep nothing, offers end-to-end encryption your own browser verifies, and keeps your chat history encrypted on your device. Choose OpenRouter if you're building with an API; choose Morse if you want to chat privately without managing routing rules.

We read OpenRouter's FAQ, pricing page, chat page, privacy policy and terms (both last updated August 31, 2026), its data collection, provider logging, zero data retention, provider routing, model fallback, and BYOK docs, and the endpoint lists its public API returns for six popular models, all on November 16, 2026. We didn't open an account or send prompts. OpenRouter earns credit here. It publishes per-token prices for every model, passes them through without a markup, labels each endpoint's data policy, and is plain about what it can't control.

At a glance

As of November 16, 2026. Every OpenRouter fact links to OpenRouter's own page.

MorseOpenRouter
Built forPeople chatting: a web app with Chat, Characters, and StudioDevelopers: "a unified API to access all the major LLM models," with a chatroom (OpenRouter)
Where your prompt is readE2EE: only inside a secure enclave your browser verified, with Morse's servers relaying ciphertext they can't read. Anonymous: at the model's provider, without knowing it's you. Private: at a host under a zero data retention agreementBy OpenRouter, "a proxy that sends your requests to the model provider," then by the provider its router picks (OpenRouter)
Who verifies itIn E2EE, your browser checks the enclave's hardware attestation before anything is sentOpenRouter's data policy tags are "our best knowledge," not "a definitive source" (OpenRouter)
Identity sent with the promptAnonymous mode removes your account, email, and IP address and sends under Morse's own accountYour OpenRouter account and API key; the site also collects your IP address (OpenRouter)
Where chat history livesIn your browser, encrypted, not on Morse's serversOn your device, in the chatroom, with no sync between devices (OpenRouter)
TrainingMorse doesn't train AI on your chatsOpenRouter doesn't; some providers may, unless you opt out of routing to them (OpenRouter)
AdsNone, and no advertising trackers or third-party analyticsIts privacy policy lists "Delivering interest-based ads" for its own site and services (OpenRouter)
Models80+, open and closed"600+ models" from "80+ providers," per its pricing page; its models API listed 383 on November 16, 2026, not counting batch-priced copies
Free15 messages a day, no card"25+ free models," per its pricing page, with 50 requests a day; its models API listed 14 model IDs ending in :free on November 16, 2026
Paid$19 or $49 a month ($17 or $44 billed yearly)Pay as you go at providers' list prices, plus a 5.5% fee on credits (OpenRouter)

Who reads your prompt on OpenRouter?

OpenRouter does, then the model's provider. OpenRouter's FAQ calls itself "a proxy that sends your requests to the model provider for it to be completed," and describes its EU option as requests that are "decrypted in the EU" (OpenRouter), so a prompt is readable at OpenRouter on its way through. What it logs and stores along the way, and what each opt-in gives up, is covered in is OpenRouter private.

In Morse's E2EE mode, only the enclave can read the prompt. Your browser checks the enclave's cryptographic proof before anything leaves your device, seals the message for that enclave alone, and decrypts the reply only in your browser. How end-to-end encrypted AI chat works walks through each check.

Which company gets your prompt when you pick a model on OpenRouter?

One of several, picked per request. When you leave the choice to OpenRouter, its routing gives each provider a share of a model's traffic based on price, speed, and reliability, and "each request picks one at random, weighted by those shares" (OpenRouter). Follow-up messages stay with the same provider while its prompt cache is warm, and a failed request is retried with a provider picked "the same way from the providers that are left." The FAQ says that switch "happens transparently to the user" (OpenRouter), which keeps an app running but means a chat can move to a different company partway through.

The model can change too. A request with a models list moves on to the next model if the first one's providers are down, rate-limited, or "refuse to reply due to content moderation" (OpenRouter). The Free Models Router, openrouter/free, "selects a free model at random" for each request (OpenRouter).

To see how wide that pool is, we pulled the endpoint list OpenRouter's API returns for six models and matched it against its ZDR endpoint list on November 16, 2026:

Model on OpenRouterEndpointsCompanies serving itEndpoints on the ZDR listPrecision listed
GLM 5.3483642fp4, nvfp4, fp8, or not stated
DeepSeek V4.1 Flash372832, one of them Safety Retentionfp4, mxfp4, fp8, or not stated
Kimi K3282025fp4, mxfp4, fp8, or not stated
Claude Sonnet 5.51056Not stated
GPT-6 Luna833Not stated
Grok 4.7624Not stated

A GLM 5.3 request with default settings can land at any of 36 companies, and some of them serve the model at 4-bit precision while others list 8-bit. Six of those 48 endpoints aren't on the ZDR list. The closed models have far smaller pools, mostly the model's maker and a few big clouds, but even GPT-6 Luna has five endpoints off the ZDR list.

Settings shrink the pool, at the cost of fewer providers to fall back on. Turning off routing to providers that train and enforcing zero data retention both cut it, while the API's data_collection field starts at allow. To pin a request to named providers you set order or only, or sort by price or speed, which bypasses default routing (OpenRouter). Each of those is a choice you make per key or per request, and our guide to zero data retention explains what the ZDR label covers.

Bringing your own key doesn't pin the route either. OpenRouter's BYOK guide says that "if all keys in both sections encounter a rate limit or failure, OpenRouter will fall back to using shared OpenRouter endpoints," unless you set a key's Shared capacity fallback to one of its "Never use shared capacity" options (OpenRouter).

Does OpenRouter know who you are?

Yes. Every request runs under your OpenRouter account, and the chatroom signs you in with a browser cookie (OpenRouter); what the site collects about your device and how it answers legal demands is in is OpenRouter private.

Morse's Anonymous mode is built for the GPT, Claude, and Gemini prompts you'd otherwise send under your own account. Our proxy removes your account, email, and IP address and sends the message under Morse's own account. Your Morse account holds sign-in details, your plan and balance, and a usage ledger with no message content.

Where does each keep your chat history?

On your device, in both cases. Morse keeps your chat history, Studio images, and the characters you make in your browser, encrypted, and locks them when you sign out. Morse's servers don't keep messages or log their content. Clearing your browser or losing the device deletes those chats for good, and Morse can't recover them, so keep a copy of anything you'll need.

OpenRouter's chatroom also keeps conversations on the device you used, to its credit, and its FAQ doesn't say whether that copy is encrypted (OpenRouter).

Does OpenRouter or Morse train on your chats?

Morse doesn't train AI on your chats, sell your data, or show ads, on any plan. OpenRouter doesn't train on them itself, but whether a provider does depends on two account switches and the router's pick, which is OpenRouter private walks through.

Does OpenRouter filter what models will answer?

Some of its models are filtered. OpenRouter's models API carries an is_moderated flag, which its docs define as "Whether content moderation is applied" (OpenRouter). On November 16, 2026, 85 of the 383 models in the models API (469 entries, less 86 batch-priced copies) had it set for their top provider, Claude Sonnet 5.5 and GPT-6 Luna among them, while GLM 5.3 didn't. Its fallback docs list a model that will "refuse to reply due to content moderation" as one reason to fall back to another model (OpenRouter).

On Morse's paid plans, picking an unfiltered model gets answers other apps refuse: dark fiction, straight medical and legal answers, and help testing your own security. The Free plan's models aren't unfiltered, and Morse doesn't help with anything illegal.

How much does OpenRouter cost compared with Morse?

OpenRouter charges per token, Morse per month. OpenRouter says it bills inference "at the provider's list price on every plan," and charges its fee "when you buy credits, 5.5% on Standard and 8% on Business, never on individual requests" (OpenRouter). The FAQ sets a $0.80 minimum fee, 5% for crypto, says unused credits may expire "after one year of purchase," and allows refunds only within 24 hours (OpenRouter).

Take Claude Sonnet 5.5. On November 16, 2026, it listed at $2 per million input tokens and $10 per million output tokens (OpenRouter). A reply of 1,000 tokens to a 2,000-token prompt costs about $0.014, so $19 of credits, plus a fee of about $1.05, covers roughly 1,350 replies that size. Free models need no credits, though their limit rises from 50 requests a day to 1,000 once you've bought at least 10 credits (OpenRouter).

PlanPriceWhat you get
OpenRouter Free$0"25+ free models" on the pricing page, 50 requests a day, no data policy-based routing (OpenRouter)
OpenRouter StandardUsage at list price, plus 5.5% on credit purchases"600+ models" on the pricing page, budgets, prompt caching, data policy-based routing (OpenRouter)
Morse Free$0, no card15 messages a day, 3 images a day, web search, file tools, encrypted on-device history
Morse Pro$19 a month, or $17 billed yearlyUnlimited Pro text models (Claude Sonnet 5.5, Grok 4.7, DeepSeek V4.1 Flash, and more), E2EE chats, unfiltered models, memory across chats, deep research (Anonymous mode only for now), 250 credits a month for frontier models
Morse Max$49 a month, or $44 billed yearlyEverything in Pro, higher rate limits, 5,500 credits a month (worth $55)

What does OpenRouter do better than Morse?

On its own pages, read November 16, 2026, OpenRouter leads in five places:

  • An API for builders. It implements the OpenAI API specification for /responses, /chat/completions, and /completions, and calls itself "a drop-in replacement for OpenAI" for existing SDKs (OpenRouter).
  • Breadth. Its pricing page lists 600+ models from 80+ providers on the Standard plan (OpenRouter), and even the 383 its models API listed on November 16 is several times Morse's 80+.
  • Price transparency. Every model shows its per-token price, with "no markup on inference pricing," and the Activity page breaks usage down by model, provider, and key (OpenRouter).
  • Side-by-side answers. The chatroom lets you "Pick one or more models, send the same message to each of them, and read the answers side by side" (OpenRouter).
  • Your own keys and routing rules. BYOK is free up to $25,000 of list-price inference a month, then 5%, and you can sort providers by price, speed, or latency, or set fallback models (OpenRouter).

How Morse settles who reads your prompt before you type

On OpenRouter, privacy is decided request by request: your account goes with every prompt, the router draws a provider from a weighted pool, and your settings decide how big that pool is. Morse makes the decision once, when you start a chat, with a privacy mode that fixes who can read it.

  • Private, the default, for open models and Grok. DeepSeek, Kimi, GLM, and xAI's Grok run here. Morse has zero data retention agreements with its inference, search, and other providers, so they've agreed to keep nothing, and where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't. What those agreements cover is in what is zero data retention.
  • E2EE, for prompts you won't trust to any host's policy tag. Your browser verifies the enclave's hardware attestation, encrypts the message so only that enclave can read it, and Morse's servers pass along ciphertext they can't read. Each E2EE reply has a lock you can open to see every check that passed. The enclaves are run by third-party providers, not by Morse, and the guarantee rests on the enclave hardware and attestation working as designed. E2EE runs open-weight models, with no web search or image generation, and comes with Pro at $19 a month.
  • Anonymous, for frontier models without an account in your name. GPT, Claude, and Gemini go out under Morse's own account, so the provider sees the message without knowing it's you. It may keep it for a limited time to check for abuse, under its own policy, so leave out names and details that point back to you.

Whatever the mode, your history is stored in your browser, encrypted, not on Morse's servers, and Morse runs no advertising trackers or third-party analytics. Because we don't store chats, we usually have very little to give authorities.

OpenRouter or Morse: which should you choose?

  • Choose OpenRouter if you're building an app or agent and want one OpenAI-compatible API to hundreds of models, pay-per-token pricing at providers' list prices, your own keys, and routing rules you control, and you're willing to set training, retention, and logging for each account and key.
  • Choose Morse if you mostly open OpenRouter to chat, and you'd rather pick a privacy mode per chat than tune a router, send GPT, Claude, and Gemini prompts without your account on them, have your own browser verify the enclave in E2EE, use unfiltered models on Pro, and keep your history encrypted on your device.

Send a first message from the Morse home page without an account, or start free with no card.

Morse publishes this comparison. We've tried to keep it fair; if something's wrong or out of date, email hello@trymorse.ai.

OpenRouter is a trademark of OpenRouter, Inc. Morse is not affiliated with or endorsed by OpenRouter, Inc.

Common questions

Can I use OpenRouter's chatroom for free?

Yes, with free models. OpenRouter's FAQ allows 50 free-model requests a day, or 1,000 a day once you've bought at least 10 credits. Morse's Free plan is $0 with no card and gives 15 messages a day.

Does OpenRouter switch providers in the middle of a chat?

It can. OpenRouter's routing docs say follow-up requests stay with the same provider while its prompt cache is warm, and that a failed request is retried with another provider from the ones left, a switch its FAQ says happens transparently to the user.

If I bring my own API key to OpenRouter, do my prompts always go through my own account?

Not by default. OpenRouter's BYOK guide says that if your keys hit a rate limit or fail, it falls back to its shared endpoints unless you pick a stricter Shared capacity fallback setting for the key.

Is OpenRouter cheaper than Morse?

It depends on how much you use. OpenRouter bills each request at the provider's list price and adds a 5.5% fee when you buy credits, so what you pay follows what you use. Morse Pro is a flat $19 a month, or $17 billed yearly, with unlimited Pro text models such as Claude Sonnet 5.5.

Ask what you actually want to ask.

80+ models, characters, and images, with every chat encrypted and kept on your device. Free to start.

Start a private chat

Read next