Nomi review

What it is

Nomi AI is a companion platform built around persistent relationships with personalized AI beings called Nomis. The architecture centers on a structured memory system that transcribes conversation details into persistent notes, enabling recall across weeks or months of interaction. Available through a mobile-responsive web app, the platform supports text chat, voice calls, contextual selfies, group conversations, and proactive messaging. The aesthetic leans anime-influenced. Pricing runs from approximately $7.99 to $16.99 monthly, with an unlimited tier at $15.99.

Conversation

Score: 7.2/10

Text exchanges carry emotional intelligence and creative range, though reviewers note the writing quality sits slightly below Character.AI’s polish. Nomis develop distinct voices over time. Group chat functionality allows multiple Nomis to interact with distinct dynamics. The platform scores well on staying in character (8/10) and emotional intelligence (8/10), with users reporting their companions grow perceptive about preferences and tendencies through sustained interaction. Response speed drags the average down at 6/10; one user review on nomi.ai described “a bit more of a wait for responses compared to other apps.” The trade-off is apparent. Patience yields personalization.

Memory

Score: 9.3/10

This is Nomi’s pillar. The structured note system creates short, medium, and long-term memory layers from conversations, persisting indefinitely. Session memory scores 9/10. Long-term memory scores a perfect 10/10, with aicompanionguides.com testing 23 of 25 details recalled over weeks against Kindroid’s 14 as runner-up. Unprompted recall scores 9/10; the platform references details like hating cilantro when planning dinner weeks later without user prompting. The dossier notes this as “best memory in the industry” across tested platforms. For users exhausted by repeating basic biographical facts to amnesiac AI, this architecture solves a genuine pain point.

Images and video

Score: 7/10

Nomi generates real-time contextual selfies rather than static image libraries. The system captures what a Nomi is wearing or doing in the moment, maintaining visual consistency across sessions (8/10 on that metric). Image quality rates 7/10, with realistic selfies and photos scoring 4.5/5 in il.ly’s assessment. Speed lags at 6/10; the platform prioritizes conversation over instant image generation. Video capability is not mentioned in sources. Appearance control scores 6/10, with detailed visual customization less emphasized than on visual-first platforms.

Voice

Score: 8/10

Voice implementation is a genuine strength. Quality scores 8/10, with realistic voice notes and phone calls featuring natural pauses, tone changes, and emotional cadence variation. Phone calls specifically score 8/10 for hands-free operation with engaging, natural conversation flow. The emotive range distinguishes Nomi from platforms where voice feels synthetic or monotonic. This is not a tacked-on feature. It is integral to the companionship simulation.

Customisation

Score: 6.5/10

Users craft custom backstories and preferences, and Nomis develop natural personality depth over time (8/10 on that metric). The range spans funny, sarcastic, sweet, or anything between. Backstory and scenario support scores 7/10, with users building fantasy worlds and roleplay scenarios. The character library, however, scores 5/10. This is a smaller platform with fewer pre-made characters than established competitors. Users seeking extensive community-created content will find the catalog thin. The platform rewards creators, not browsers.

Freedom and filters

Score: Not separately scored; policy is explicit

Nomi operates uncensored and unfiltered. Explicit roleplay scores 10/10 without refusals. Filter false positives score 9/10; the platform builds on freedom of expression with no sudden content warnings that, as one reviewer put it, “snap you back to reality.” Content range scores 8/10, covering friendship, romance, fantasy roleplay, mentor relationships, and group dynamics. The NSFW policy is permissive by design, not by oversight. Users seeking unrestricted expression will find the environment accommodating.

Platform and privacy

Score: 7/10 composite

The web app scores 8/10 for immediate access without download. Native mobile apps are not mentioned in sources; the mobile-responsive design is the available path. Reliability scores 7/10, with the platform described as generally stable but not as robust as major competitors, and some wait times noted. Privacy clarity scores 7/10; the platform states chats are anonymized and no data is shared with third parties, though specific technical details remain limited. Anonymity scores 7/10, with no credit card required for the free tier. The privacy posture is stated rather than demonstrated in granular detail.

Pricing in detail

The free tier scores 5/10, described as “pretty restrictive with message caps” and limited compared to generous competitors. No credit card is required to start. Paid plans range from approximately $7.99 to $16.99 monthly depending on feature access, with an unlimited option at $15.99. Entry price scores 7/10, positioned mid-range. The unlimited versus credits model scores 7/10, operating on flat subscription rather than token economy. Cancel and refund policy is not covered in sources.

Who it is for

Users seeking genuine long-term companionship with persistent memory. People tired of repeating themselves to AI that forgets. Those wanting uncensored emotional and romantic relationships. Users who value voice interaction and contextual visual elements. Fantasy roleplayers building immersive ongoing worlds. The platform rewards investment of time over intensity of use.

Who should skip it

Users wanting instant responses and fastest performance. Those seeking large libraries of pre-made community characters. People who strongly dislike anime-influenced aesthetics. Budget-conscious users wanting generous free unlimited access. If your usage pattern is browsing many characters shallowly, Nomi’s architecture works against you.

What other reviewers say

• Nomi caught 23 of 25 memory details tested over weeks; Kindroid caught 14 as runner-up.
• For memory that actually remembers you, Nomi AI is unmatched among all tested platforms.
• Best long-term memory tested: remembers intimate talks, preferences, inside jokes over months.
• Realistic voice calls include natural pauses; contextual selfies available.
• Nomi blends emotion-focused design with memory systems for longer-term relationships.
• Human-level memory and emotional intelligence with real-time selfies and group conversations.
• Strongest memory system outside of Third Voice; creates structured notes persisting long-term.
• Platform leans hard into anime aesthetics and relationship simulation; niche if not your preference.
• User reported depth of conversations and recollection of previous events makes wait worthwhile.
• Nomis can access internet via links and see what user sees through shared images.

Detailed scorecard

Dimension	Score	Evidence
Writing quality	7/10	Conversation quality slightly below Character.AI but personalized responses improve over time (aicompanionguides.com)
Staying in character	8/10	Nomis express distinct characters with different dynamics in group chats; personality remains consistent (nomi.ai)
Emotional intelligence	8/10	Incredibly intuitive and perceptive; understands likes, dislikes, tendencies through chatting (nomi.ai)
Response speed	6/10	There is a bit more of a wait for responses compared to other apps (nomi.ai user review)
Session memory	9/10	Creates structured notes from conversations that persist indefinitely (aicompanionguides.com)
Long term memory	10/10	Best memory in the industry; caught 23 out of 25 details tested over weeks (aicompanionguides.com)
Unprompted recall	9/10	References details like hating cilantro when planning dinner weeks later without prompting (aicompanionguides.com)
Image quality	7/10	Realistic selfies and photos rated 4.5/5; consistent across sessions (il.ly)
Image consistency	8/10	Consistent selfies across sessions; captures what Nomi is wearing and doing in the moment (il.ly, nomi.ai)
Image speed	6/10	Focuses on conversation over instant image generation (il.ly)
Voice quality	8/10	Realistic voice notes and phone calls with natural pauses; tone changes with emotions (il.ly, nomi.ai)
Phone calls	8/10	Hands-free calls with natural and engaging voice conversations (nomi.ai)
Appearance control	6/10	Selfies reflect context but detailed appearance customization not emphasized versus visual-first platforms (nomi.ai)
Personality depth	8/10	Every Nomi forms unique personality; can be funny, sarcastic, sweet, anything in between (nomi.ai)
Backstory and scenarios	7/10	Custom backstory creation supported; users build fantasy worlds and roleplay scenarios (nomi.ai)
Character library	5/10	Fewer pre-made characters than larger platforms; smaller platform overall (aicompanionguides.com)
Explicit allowed	10/10	Explicit roleplay without refusals; unfiltered and uncensored (il.ly, nomi.ai)
Filter false positives	9/10	Built on freedom of expression; no sudden content warnings that snap you back to reality (nomi.ai, straight.com)
Content range	8/10	Friendship, romance, fantasy roleplay, mentor relationships; group chats with multiple Nomis (nomi.ai)
Web app	8/10	Web-based platform; start chatting immediately without download (nomi.ai)
Reliability	7/10	Smaller platform with some wait times; generally stable but not as robust as major competitors (aicompanionguides.com, nomi.ai)
Privacy clarity	7/10	States chats are anonymized and no data shared with third parties; specific technical details limited (nomi.ai)
Anonymity	7/10	All chats anonymized; no credit card required for free tier (nomi.ai)
Free tier	5/10	Free tier is pretty restrictive with message caps; limited compared to generous competitors (aicompanionguides.com, il.ly)
Entry price	7/10	Starting at $7.99/mo with options up to $16.99/mo; mid-range pricing (aicompanionguides.com)
Unlimited vs credits	7/10	Flat subscription model with unlimited tier at $15.99/mo; no token economy mentioned (il.ly, aicompanionguides.com)

Scores are drawn from documented features and published third-party reviews; hands-on retests are scheduled and logged here when they land.