Skip to content

Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.

fix(avatar): reword four em-dash comments so the sign() docs and mood table pass the push lint

nirholas683e5c8ce6de
diff --git a/src/agent-avatar.js b/src/agent-avatar.js
index 4d78b2c95..8715118c0 100644
--- a/src/agent-avatar.js
+++ b/src/agent-avatar.js
@@ -57,7 +57,7 @@ const DECAY = {
patience: 0.035, // half-life ~20s — sustained waiting state
curiosity: 0.12, // half-life ~8s — alert, engaged
empathy: 0.055, // half-life ~13s — slow to fade, like real empathy
- uncertain: 0.1, // half-life ~7s — hedged speech signal
+ uncertain: 0.1, // half-life ~7s: hedged speech signal
};
// Vocabulary scored for emotional valence
diff --git a/src/element.js b/src/element.js
index 52ae69546..b57bec414 100644
--- a/src/element.js
+++ b/src/element.js
@@ -1113,7 +1113,7 @@ class Agent3DElement extends HTMLElement {
});
})
.catch(() => {
- /* optional affordance — never break the viewer */
+ /* optional affordance, never break the viewer */
});
}
@@ -2956,8 +2956,8 @@ class Agent3DElement extends HTMLElement {
/**
* Perform `text` in American Sign Language on the avatar, without sending it
* to a brain. Turns the engine on if the `sign-language` attribute has not
- * already, and resolves with `{ signed, spelled }` — which words came from
- * the lexicon and which fingerspelled — or `null` when the loaded rig has no
+ * already, and resolves with `{ signed, spelled }` (which words came from
+ * the lexicon and which were fingerspelled), or `null` when the loaded rig has no
* finger bones to sign with.
*
* @param {string} text