Overview
HammerAI lets users chat and roleplay with AI characters through two routes: locally-run LLMs (in-browser via WebGPU, or the desktop app with Ollama/llama.cpp and GPU acceleration — fully offline) or cloud-hosted models on HammerAI's servers. The official site describes it as 100% free to chat on both local models and cloud-hosted models, and no account or login is required to start; local-mode messages stay on the user's device and conversations are deleted when the app is closed.
Beyond one-on-one chat, the platform offers AI story writing, AI image generation, group chats with up to ten characters at once, lorebooks (dynamic memory), personas, and text-to-speech (746 voices, 34 languages). Advanced users can proxy to any external LLM API and tune model samplers and parameters (Min-P, Top-P, Top-K, temperature, repetition penalty). Official pricing-page wording: "HammerAI is built by one developer as a passion project"; subscriptions are processed through GankNow.
Pricing is official and credit-based. Monthly plans: Free ($0) with 600 cloud messages/day, 8,192-token context, 2 cloud LLMs; Starter ($9/month) with 1,500 credits/month, unlimited cloud messages, 16,384-token context, 4 cloud LLMs; Advanced ($18/month) with 3,500 credits/month, 32,768-token context, 7 cloud LLMs; Ultimate ($35/month) with 9,000 credits/month, 65,536-token context, 14 cloud LLMs. One-time credit packs run from $1.99 (200 credits) to $99.99 (20,000 credits); purchased packs never expire, while subscription credits expire each billing period. Local-model chat is free and unlimited on all plans.
Key Features
- AI character chat & roleplay
- AI story writing
- AI image generation (26 image models)
- group chats (up to 10 characters)
- lorebooks/dynamic memory
- personas
- text-to-speech (746 voices, 34 languages)
- local LLMs via desktop (Ollama/llama.cpp, offline)
- in-browser WebGPU inference
- proxy to any LLM API
- advanced samplers
- rolling-summary conversation memory
- no login required
- no ads on desktop
