selan.ai Gateway Reactor

Compare · OpenRouter and Selan

Is Selan an OpenRouter alternative for a team's coding tools?

Not quite: Selan uses OpenRouter as one of its providers, so a team can run both. OpenRouter sells 500+ models per token, from prepaid credits or your own provider keys. Selan runs a team's Claude Code, Codex and Claude Desktop on the Claude and ChatGPT subscriptions you already pay for, and reaches OpenRouter only past them. If every request is billed per token anyway, OpenRouter alone is the better answer.

subscription openrouter

How do OpenRouter and Selan compare?

OpenRouterSelan
What it isA hosted API in front of 500+ models from 80+ providers, in the OpenAI format and Anthropic'sA hosted gateway, with a CLI that launches Claude Code, Codex and OpenCode through it
What it runs onCredits you buy, or your own provider API keys (BYOK)Your Claude, ChatGPT and Ollama Cloud subscriptions first, then your API keys, OpenRouter's among them
A developer's credentialAn OpenRouter API key set as ANTHROPIC_AUTH_TOKEN, or a browser sign-in through ori claudeOne sign-in with Google or Microsoft through selan login; no key on the laptop
ClientsAny app on its API; Ori Harness launches Claude Code, Codex, OpenCode and other agent CLIsClaude Code, Codex, OpenCode and Claude Desktop
BudgetsPer member and per key, daily, weekly or monthly, through guardrails; credit spend only unless BYOK spend is switched inPer person, by day, week or month in UTC, on pay-per-use keys
Sensitive dataGuardrails that redact or block personal information and your own patternsSecret scanning, off until an owner turns it on
What you pay itA fee of 5.5% on credit purchases (8% on Business); BYOK free up to $25,000 of list-price inference a month, then 5%A flat monthly plan sized by seats, free for up to three; no share of tokens, since your provider bills you directly

OpenRouter sells access to models. Selan spends the subscriptions you already own first, then your OpenRouter key.

When is OpenRouter alone the better answer?

When your team pays per token for every request, and holds no Claude or ChatGPT seats to spend first. OpenRouter then gives one bill across 500+ models and fails over between providers that host the same model.

Its Ori Harness signs a developer in through the browser and runs Claude Code, Codex or another agent CLI on the organization's models, guardrails and workspace budgets. That covers more agents than Selan does.

It also suits an application of your own. Its API is OpenAI-compatible, Zero Data Retention routing is on every plan, and the Business plan adds EU in-region routing.

When is Selan the better answer?

When your team already pays for Claude or ChatGPT subscriptions. Selan sends each request to a subscription first and falls through to a key billed per token only when the seats are capped. It reads each Claude account's 5-hour and 7-day windows from every response.

OpenRouter's documentation describes credits and provider API keys, and no way to run on a Claude or ChatGPT subscription. Through it alone, every Claude Code session is billed per token.

Selan records Claude Code, Codex and Claude Desktop under each person's name, and the owner's Usage page ranks people and teams. The trade-off is that requests travel through Selan, so if Selan is unavailable those requests fail.

How does Selan use OpenRouter?

An owner connects the company's own OpenRouter key in Selan. A developer then names any model OpenRouter serves, with the provider before the first colon:

selan claude --model openrouter:moonshotai/kimi-k3

Everything after the colon is OpenRouter's own model id, passed through as it is. Selan also falls back to OpenRouter on its own: for GPT when the ChatGPT plan is capped, and for the open-weight models when an Ollama Cloud plan runs out.

OpenRouter bills those requests to your account, at its prices. Selan records the cost OpenRouter reports for each request under the person who made it. A person's spend limit counts it, because the key is billed per token.

The routes are in Providers and models, and naming a model in Run any model.

Good to know: OpenRouter's own guide says Claude Code is optimized for Anthropic models and may not work correctly with other providers. Anthropic doesn't support routing Claude Code to non-Claude models through any gateway, Selan included.

How do you move Claude Code from OpenRouter to Selan?

OpenRouter's setup points ANTHROPIC_BASE_URL at https://openrouter.ai/api, puts the key in ANTHROPIC_AUTH_TOKEN and blanks ANTHROPIC_API_KEY. selan claude sets the base URL and token for the session it launches, in place of whatever the shell exported.

Connect the same OpenRouter key in Selan and the same models stay one flag away. Limits per person are on the spend page.

Start for free

Free for up to three seats, and the Free plan needs no card.

The short version

What OpenRouter does is as its pricing page, its Claude Code, Ori Harness, guardrails, workspaces and BYOK documentation at openrouter.ai described it, read on 21 October 2026. What Selan does is selan-cli and selan-control as they stood on the same day.