Site overview / Mirasim

Mirasim

Multi-agent workspace with a paid Go plan; connecting your own accounts or API keys is free, but third-party usage is not.

Open service / register →

First-day credits / plan

First-day credits / plan: Go · $1/month
Status: Reachable
Snapshot updated: 2026-10-06 06:23 UTC

Estimates per 5-hour window

Kimi K3: ≈130 requests

GLM 5.3 Flash: ≈1,900 requests

DS 4.1 Flash: ≈7,800 requests

Paid plan, excluded from free-credit totals. These are provider estimates from one shared allowance: do not add model counts or treat them as guaranteed daily/monthly totals.

Official pricing · Manually checked: 2026-09-23

Credit policies are manually recorded from provider notices; public endpoints do not always expose them. Models, pricing and reachability are refreshed automatically where permitted. Reachability does not prove successful signup or model access.

Registration requirements

Use the full referral link and sign in with GitHub or an email code. No public rule promises a signup bonus to the invitee. Install the official desktop client or follow its CLI documentation; no general-purpose relay base URL has been publicly verified.

Rewards, limits and important caveats

Go usage estimates share one allowance and cannot be added together. The inviter earns one Go month only after three invited friends all subscribe to Go. Claude/Codex Basic, Pro and Max plans are separate from Go. Alpha software may be slow or interrupted; old beta compensation is not a new-user benefit. Robots rules permit probing the pricing page, not the referral path; online status does not verify signup, payment or inference.

Client setup

Create an API key in the service dashboard and use only a model listed for your account. Anthropic base URLs do not include /v1; OpenAI-compatible URLs usually do. A protocol endpoint does not guarantee access to every model or client.

Dashboard

Provider documentation (may be Chinese)

Public model snapshot

Only models returned by the public snapshot are listed; missing data does not mean there are no models. Confirm current prices, account groups and availability in the dashboard. Per-request prices are not token prices.

Not publicly specified