SoulkynFreemium
Uncensored AI companion platform with customizable personas, NSFW image/video generation, voice chat, and memory driven conversations. True creative freedom in one seamless platform.
uncensored AI chatcharacter customization
Uncensored character chat with lorebooks, ten character group chats and local models that run entirely on your own machine.
HammerAI is the rare character platform that can move an entire roleplay session onto your own computer: the desktop app runs local models through Ollama, stores chats on the device, and keeps working offline. Lorebooks, ten-character group chats, editable samplers, and context up to 65,536 tokens give it unusual depth, but the web and cloud side is less tidy because plan cards, comparison tables, and an older FAQ publish conflicting model and credit counts.
HammerAI splits into two distinct products behind one interface. The desktop app can run a local model, save characters and conversations on the machine, and operate without a connection. The web app and cloud models remove the hardware requirement, while paid credits add hosted image generation and text-to-speech.
Editor's Pick
Voice, video & uncensored chat — 100% private. Your perfect partner, built to your exact specifications. Free to start.
That choice changes the risk and the experience. Local mode gives the most privacy and control but depends on the computer and model you supply. Cloud mode is easier on a phone or modest laptop, yet it brings daily limits on the free plan and a pricing page whose model counts are not internally consistent.
HammerAI is an uncensored character-chat and interactive-story platform for web, Android, iOS, macOS, Windows, and Linux. It supports community characters, custom cards, user personas, lorebooks, group conversations, local and cloud language models, image generation, and spoken replies.
The local route is its defining mechanic. Ollama is bundled with the desktop app, more than twenty models work without manual configuration, and an Ollama Modelfile can bring in another compatible model. WebLLM also enables browser-local inference on a WebGPU-compatible browser.
HammerAI fits writers and roleplayers who want to own the character files, tune generation settings, and keep sensitive conversations off a hosted chat service. It also suits longer world-building projects where a lorebook and rolling summaries matter more than a polished companion-app onboarding flow.
A lightweight hosted companion is simpler when local model setup sounds like work. HammerAI can supply cloud models, but its sharpest advantage appears only when the desktop app and local inference are part of the plan.
| Feature | What it does | User benefit | Limitation or note |
|---|---|---|---|
| Local model mode | Runs language models through bundled Ollama on desktop or WebLLM in a compatible browser. | Keeps the model and conversation on the device and allows offline sessions after setup. | Speed and response quality depend on the selected model and available hardware. |
| Character cards and personas | Separates personality, scenario, example dialogue, opening message, user persona, and generation settings. | Makes a character portable and gives the author direct control over how a scene starts and responds. | The amount of control is closer to a roleplay workbench than a one-click companion builder. |
| Lorebooks | Injects world facts when matching terms appear instead of keeping every fact in the active prompt. | Preserves setting continuity without spending the full context window on background material. | Entries still need careful keywords and writing to activate at the right moment. |
| Group chat | Places up to ten characters in one conversation. | Supports ensemble scenes, party dialogue, and conflicts that a one-companion app cannot stage. | Large casts put more pressure on context and character-card quality. |
| Context and rolling summaries | Publishes context limits from 8,192 tokens free to 65,536 on Ultimate, with automatic conversation summaries on paid tiers. | Keeps longer story arcs usable after the raw transcript grows beyond the active window. | The price table and feature comparison disagree about paid cloud-model counts. |
| Cloud images and speech | Paid access covers 22 image models and text-to-speech with 746 voices across 34 languages. | Adds a visual or spoken beat without moving a character into another service. | Images cost 10 credits and speech costs 5; the free plan receives no credits. |
The desktop build packages Ollama and can run a conversation without contacting a cloud model. Characters, stories, and chats remain on the computer, which is materially different from a hosted platform that merely promises not to train on messages.
Local operation does not make every HammerAI session private by default. Selecting a HammerAI cloud model or proxying to another provider sends the request off the device under that service's rules. The privacy advantage therefore depends on which model route is active.
Character creation exposes personality, scenario, example dialogue, first message, and user persona rather than reducing a companion to a few appearance choices. Temperature, repetition penalty, Top-P, Top-K, and the system prompt can be adjusted per character.
Lorebooks add a second layer for recurring places, factions, rules, and supporting characters. Entries surface when their keywords appear, so a large world can stay available without occupying the prompt during every exchange.
A group can hold ten characters, enough for a full party, family, or ensemble scene. This is more than placing several avatars in a room: each card brings its own definitions and dialogue examples into the shared exchange.
Context scales from 8,192 tokens on Free to 16,384 on Starter, 32,768 on Advanced, and 65,536 on Ultimate. Paid rolling summaries compress older events, while local users can choose a model and context configuration their hardware can sustain.
HammerAI's hosted image tool offers 22 models and charges 10 credits for a generation. It can illustrate a character or scene alongside chat, while the documentation also covers external Forge and ComfyUI workflows for people who prefer their own image stack.
The free plan includes chat but no media credits. A $9 Starter allowance of 1,500 credits equates to as many as 150 images only when none of those credits are spent on speech.
HammerAI Image Generator
Cloud text-to-speech turns a reply into audio for 5 credits, with 746 voices and 34 languages listed across the paid plans. It is a reply-reading feature rather than a live telephone-style call.
The shared wallet creates a direct tradeoff: every spoken reply uses half the credits of an image. Purchased credit packs remain available after a billing period, whereas the monthly subscription allowance expires.
Free includes all local models, API proxying, two cloud models, 600 cloud messages per day, lorebooks, group chat, and an 8,192-token context ceiling. It does not include credits for hosted images or speech.
Starter costs $9 per month with 1,500 credits and 16,384 context. Advanced costs $18 with 3,500 credits and 32,768 context. Ultimate costs $35 with 9,000 credits and 65,536 context. Paid chat is unlimited, an image costs 10 credits, and a speech generation costs 5.
The tier cards currently list 4, 7, and 13 cloud models, while the comparison table lists 4, 6, and 10. A pricing FAQ also retains older monthly credit figures. Treat the checkout and account entitlement as decisive until those sections agree.
| Tool | Best for | Key advantage | Main limitation | Review or comparison link |
|---|---|---|---|---|
| Soulkyn | A hosted companion with multimedia, voice, and a conventional account experience | HammerAI offers true offline inference, lorebooks, and direct sampler controls. | Soulkyn packages chat and companion media more smoothly for people who do not want to manage local models. | Soulkyn review |
| Chub AI | A large portable character-card and lorebook ecosystem | HammerAI can run cards locally and accepts ordinary payment methods for its cloud plans. | Chub's single subscription includes its largest hosted model, voice, and multimedia without separate media metering. | Chub AI review |
| JANITOR.AI | Free hosted community-character chat with optional external model keys | HammerAI provides an offline desktop route, lorebooks, and locally stored files. | Janitor AI is quicker to enter when a browser-based community library matters more than local control. | Janitor AI review |
Chub AI is the closest comparison for character cards and lorebooks, while Janitor AI and SpicyChat AI reduce setup for hosted community roleplay. Soulkyn moves further toward a polished multimedia companion with voice and video.
Story AI by DreamGen is the writing-focused option when generating and revising narrative matters more than simulating a cast. HammerAI remains distinctive when offline ownership, group scenes, and low-level model control belong in the same workspace.
HammerAI earns its place through architecture rather than surface polish. Running a local model, retaining portable JSON data, and keeping lorebooks and ten-character scenes on the device gives serious roleplayers control that hosted companion apps rarely expose.
The free plan is genuinely useful for local chat and limited cloud messaging. Hosted images, speech, larger context, and unlimited cloud messages make the paid tiers relevant, but the contradictory plan counts deserve attention before subscribing.
Choose HammerAI when ownership and configurability outweigh convenience. Choose a hosted alternative when the priority is immediate companion media, simpler onboarding, or one clearly bounded subscription.
Limited Access
Voice, video & uncensored chat — 100% private. Your perfect partner, built to your exact specifications. Free to start.
Play the choice-driven relationship sim.
Build a custom companion from scratch.
The jealousy-driven sim experience.
Daily messages, photos, and personal attention.
Compare every AI girlfriend app.
2026 honest comparison rankings.
Uncensored AI companion platform with customizable personas, NSFW image/video generation, voice chat, and memory driven conversations. True creative freedom in one seamless platform.
Adult AI companion platform with unlimited free text, persistent memory, custom characters, requested images, voice, and premium video generation.
AI dominatrix chatbot platform for interactive BDSM roleplay with customizable characters and image generation.
Media first AI girlfriend platform for fictional adult companions, private chat, memory, AI images, short V5 video, live rooms, and character consistent NSFW roleplay.
AICupid combines adult character chat, character creation and import, AI images, short character videos, TTS, speech input, and voice calls in one browser platform.
Uncensored NSFW AI companion platform with customizable virtual girls, roleplay, and integrated image generation for immersive adult experiences.
Adult anime art studio with Anima, Illustrious and Pony models, an OC Maker, image editing, upscaling, and Seedance text to video and image to video with audio.
Polished AI girlfriend and companion platform with chat, voice, images, video style responses, character creation, memory, and premium NSFW roleplay.
Multimodal AI character platform with text, voice TTS, and image generation in one interface, huge in South Korea with 40,000+ community characters and a creator monetization system.
An adults only AI companion platform combining community character chat, custom anime or realistic companions, image generation, image to video, voice calls, scenarios, and personas.
Yes, with real limits. The free plan carries 600 cloud messages a day at 8,192 tokens of context across two cloud models, plus unlimited use of any local model through the desktop app. It receives no credits, so cloud image generation and text to speech are not available until you pay.
The desktop app can. It loads local models through Ollama and works completely offline, storing characters, conversations and stories as JSON files on your own computer. Cloud models, image generation and text to speech all require a connection.
Its privacy policy states it does not collect information about conversations with characters and does not use user data to train models internally. It is also explicit that messages sent to third party model providers may be stored or used by those providers under their own policies, which running a local model avoids entirely.
Credits pay for cloud image generation at 10 each and text to speech at 5 each. Cloud chat messaging is unlimited on every paid plan and does not consume them. Monthly credits expire at the end of the billing period, but separately purchased credit packs never expire.
No. The vendor states that all purchases are final because its payment processor does not support refunds. Subscriptions can be cancelled at any time through the payment account, or through device settings for mobile subscriptions.
Answer 6 quick questions and get matched to the platforms that fit your preferences.
Find Your AI Companion
