Features
A private workspace that does the work.
TalkNova AI is one managed account for many AI models, with your files, your projects and your conversation history under your control.
Every model, one picker
52 verified text models — Claude Opus 5.5, DeepSeek V4.1 Flash, GLM 5.3 Flash, Kimi K3, Nova Muse, Nova Pulse, Claude Opus 4.8, Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.2 Astra, GPT-6.2 Sol, GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna, GPT-6 Luna Pro, GPT-6.1 Sol Pro, GPT-6 Astra Pro, Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro 0813, DeepSeek V4.1 Flash 4, DeepSeek V4 Flash, DeepSeek V4.1 Flash (Uncensored), GLM 5.3, GLM 5.2, GLM 5.3 (Uncensored), GLM 5.3 Flash (Uncensored), Kimi K2.5, Nova Vex (Uncensored), GPT OSS 120B, Gemma 4, Gemma 4 31B, Qwen3.5 9B, Qwen3.8 27B (Uncensored), Qwen3.5 122B, Qwen3, Qwen3.6 35B, Qwen 4 Max, Qwen3.6 27B, Qwen3.8 27B, Qwen3.8 2.4T, Qwen3.8 Flash, Nemotron 3.5 Lightning, NVIDIA Nemotron 3 Ultra 550B, Muse Spark 816B, Muse Glimmer 30B, Muse Glimmer 30B, Llama 3.3 70B Instruct, Llama 3.1 8B Instant, MiniMax M3.1 Flash, Inkling Small and Grok 4.5 — in a searchable picker. Routes that stop answering are removed after a model check rather than left selectable.
Work with your own files
Attach text and source files or a ZIP archive and ask questions about them. Archives are read in memory and never extracted; binaries, dependencies and private key or .env files are skipped with a visible notice.
Readable code and HTML previews
Answers get syntax highlighting, per-block copy and download, and complete file bundles you can save as a ZIP. HTML answers open in a sandboxed preview with network requests blocked.
Chats follow your account
Chat history is saved on your device and synced to your account on the web, desktop and Android, so the same chats appear everywhere you sign in. Sync is always on; you can delete any or all conversations. Saved personalization, Agents and video jobs have separate account-bound server storage.
Account controls you can see
Sign in with Google, check your plan and remaining monthly usage from the sidebar, and manage or regenerate a personal API key for OpenAI-compatible clients.
Web, Mac, Windows and Android
The desktop and Android apps share your account, models and allowances. Desktop adds linked project folders with reviewed file edits, and every app installs signed updates.
Image generation
Text-to-image generation in web chat has its own monthly plan allowance and needs a configured official credential to answer.
Private by construction
Provider credentials stay on the server, the browser only ever calls TalkNova, prompts and responses are not logged, and production builds ship without source maps.
See the live model catalog
Availability and usage limits vary by model and plan. The model directory always reflects what is configured right now.