At a glance

Model IDlucidityai/synth-2.5-flash
Free ID-
Made byLuciditySH
Context window30,000 tokens
Longest replyNot published
InputText
ReasoningNot listed
Tool callingNot listed

Synth 2.5 Flash Preview is LuciditySH's creative writing model. Every account can call it from day one, with no credits, no plan and no key from another provider, and it has its own daily allowance that referrals, invites and Discord boosts can raise. That makes it the usual first stop for a new account: connect a chat app, pick Synth, and check that your character card and preset behave before you decide whether to pay for anything.

What Synth is for

Synth was built for story and character writing. It is not listed with image input, and the catalog does not list reasoning effort or tool calling for it, so use it for text chat and keep agent or coding work on the models that list those features.

Every community preset, lorebook, prompt and plugin on Composite works with Synth. If you found a setup in Discovery, attach it to your key and run it on Synth to see how it reads before trying it on a paid model.

Working inside 30,000 tokens

Synth's context window is 30,000 tokens. That holds a detailed character card, a preset and a fair amount of chat history, but a long campaign will outgrow it. Once a conversation passes the window, Composite trims the oldest messages so the request still fits, and the model stops seeing those early scenes.

A few habits keep long Synth chats coherent:

  • Set your chat app's context size to 30,000 or a little less, so the app trims in a predictable place instead of sending more than fits.
  • Keep the character card and preset short. Every token they use is a token of history the model loses.
  • Move facts that only matter sometimes into a lorebook, so they are sent when a keyword comes up rather than on every message.
  • Paid accounts can turn on a premium memory layer on the key, which keeps older history available to the model. The features guide lists which layers each account can use.

The model ID

Use lucidityai/synth-2.5-flash. Synth has no separate :free ID because the plain ID is already free. Older setups that still send lucidityai/synth-2.5-flash:free keep working; Composite maps that name to Synth.

Connect it to your chat app

  1. Create an API key on the API Keys page.
  2. In SillyTavern, ChubAI, Agnai or RisuAI, choose an OpenAI-compatible provider and set the base URL to https://composite.lucidity.sh/v1. JanitorAI wants the full URL: https://composite.lucidity.sh/v1/chat/completions.
  3. Paste the key, then type lucidityai/synth-2.5-flash as the model name.
  4. Send a short message to check the connection before you load a long chat.

Step-by-step setup for each app is in the JanitorAI and SillyTavern guides, and the docs hub covers the others.

From code, it is a normal OpenAI-compatible request:

curl https://composite.lucidity.sh/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "lucidityai/synth-2.5-flash", "messages": [{"role": "user", "content": "Hello"}]}'

Who gets Synth, and how much

Any account can use Synth with no provider key: 80 requests a day, or 300 a day for the first week after joining with a referral code. Inviting others raises your own cap: 10 a day for each friend who signs up with your link and uses Synth, 5 for each person who joins the Discord server through your invite and connects their account, and 100 while you boost the Discord server, up to 200 a day. Any credits or an active plan make Synth unlimited.

Synth has its own counter. It does not use up the 80 requests a day shared by the other free models, and they do not use up Synth's. The count resets once a day at the time shown on your account page. The features guide keeps the current rules.

More