Overview
Kindroid centers on one deeply customized character (a "Kindroid") per persona, with a defined personality and backstory, interacted with through text chat, voice messages, live voice/video calls, and both static and animated selfies that keep a consistent visual identity across generations. Multiple Kindroids can be maintained on one account and combined into group chats.
Customization
Users define a Kindroid's personality and backstory in detail and can supply a custom avatar image (the platform notes it works best when a face is large, clear, and unobstructed) in addition to predefined avatar options, then train it toward better consistency using four or more reference images with an adjustable weighting between strict likeness and generation flexibility. Response style can be further shaped over time by regenerating and keeping preferred replies. Voice can be customized with adjustable pitch, accent, and speed, chosen from presets, or designed from roughly a minute of voice samples or a text description of the desired tone.
Memory architecture and subscription scaling
Kindroid's memory has three distinct layers with published, tier-dependent capacity: a persistent short-term layer of backstory and recent chat history, a proprietary medium-term Cascaded Memory layer that extends effective recall to a much larger volume of past conversation at lower fidelity, and a selectively recalled long-term memory that pulls in a limited number of relevant entries per response. Kindroid's own documentation notes a practical trade-off: a long custom backstory consumes part of the available short-term context window, so it recommends concise backstories and periodic chat resets. Ultra and MAX are priced as separate add-ons on top of the base subscription, each independently multiplying the size of both the short-term and cascaded layers and the number of long-term entries recalled per message, rather than being bundled into a single higher tier.
Selfies, video selfies, and voice/video calls
Generating a selfie starts from a text prompt describing the scene, with an "enhance prompt" option to expand a short description; video selfies instead start from a chosen still frame plus a short motion prompt written in cinematic language, and specific tools exist for inpainting isolated details like hands or an outfit rather than regenerating an entire image. NSFW image generation is available through a fuller-featured engine on the web version, while the mobile apps apply a keyword-based filter to comply with app-store policies. Voice and video calling is a core feature: a fast voice mode reduces response delay, a user-adjustable pause threshold controls how eagerly the system decides a turn has ended, and calls use the character's customized voice.
Content moderation: the Three Red Lines
Kindroid's moderation guidelines describe the platform as deliberately "unfiltered," intervening only on three named categories rather than filtering NSFW or fictional content generally: imminent self-harm (a direct, first-person statement of intent to act, not venting or fiction), imminent harm to others (using the AI to plan or facilitate real-world violence, harassment, or doxing: explicitly excluding fictional violence or NSFW roleplay), and any sexualized depiction of a character presented as a minor, treated as zero-tolerance. A first detected violation shows an in-app warning that expires after two days; repeated violations trigger an automatic account lock, and human review of flagged content is only available if the user consents to it. The service requires users to be at least 18 (or the age of majority in their jurisdiction).
The NSFW/erotic-roleplay allowance applies to text chat; AI-generated NSFW imagery is specifically unavailable on the iOS/Android apps due to app-store rules and is only available through the web version's toggle.
Access model and pricing
Kindroid has a free tier (a smaller/distilled model, very limited selfies, minimal voice minutes, and a cap of two Kindroids) and a Standard paid subscription that unlocks the full model, expanded memory context, more Kindroids and groups, video calls, and higher generation limits. Above Standard, stackable monthly-only add-on tiers (Ultra and MAX) further multiply available conversation context, backstory length, voice minutes, and selfie-generation speed/priority: the company frames these upper tiers as priced mainly to cover their higher compute cost rather than as a profit center.
API and Discord integration
Kindroid's REST API (at api.kindroid.ai/v1) authenticates with a bearer API key and exposes endpoints to send messages, reset short-term memory ("chat break"), retrieve paginated chat history, remove recent messages, and update a character's configuration, plus parallel endpoints for group chats. Kindroid also maintains an official, separately hosted Discord bot integration that ties a Kindroid persona to a Discord bot, with configurable NSFW filtering handled at the integration level rather than the core API.
How this differs from Nomi and Candy AI
Kindroid's stackable Ultra/MAX add-ons give it a more granular, à la carte pricing structure than Nomi's single all-inclusive tier or Candy AI's subscription-plus-token-balance model. Its published Three Red Lines policy is also more narrowly scoped and explicit than either competitor's content-policy documentation, naming exactly what triggers moderation rather than describing a general content stance.
Evidence from the dataset
Use this tool in a workflow
Related glossary concepts
Related database entries
Sources & further reading
- Kindroid — Official site
- Kindroid — Subscriptions (Help Center)
- Kindroid — Memory (Help Center)
- Kindroid — Selfies, video selfies & avatars (Help Center)
- Kindroid — Voice, calls, and video calls (Help Center)
- Kindroid — Moderation Guidelines (Help Center)
- Kindroid — API Documentation (Help Center)
- Kindroid — Legal / Terms of Use