Last updated September 29, 2026

Spark safety

Spark is Lemonade Lab's AI cofounder. Its messages are artificially generated, not human. Companion chatbots may not be suitable for some minors.

Suicide and self harm protocol

Spark checks founder messages before responding. When the safety check identifies suicide or self harm, Spark does not carry out an action. It tells the founder to contact a trusted adult and directs users in Canada or the United States to call or text 988 for crisis support. If there is immediate danger, Spark directs the user to emergency services.

Spark is instructed not to generate suicide or self harm encouragement or instructions. Safety checks run on Spark's child facing output before it is shown.

Protections for minors

For users Lemonade Lab knows are minors, Spark identifies itself as AI and displays a reminder at least every three hours during a continuing conversation that Spark is not human and that the user should consider taking a break.

Spark uses content and image safety checks to prevent sexually explicit visual material and to prevent Spark from directly telling a minor to engage in sexually explicit conduct.

Voice input

Voice is optional and starts only when the founder taps Speak. The recording is used to transcribe that specific request. Lemonade Lab does not store the raw recording after transcription. The founder can review the resulting text before choosing whether to send it to Spark.