Knowledge base / Money and extras

Nymbot

An AI assistant built into the app. Free in public channels, paid in a private encrypted chat, and able to read a git repository if you connect one.

Nymbot in channels

Anything beginning with ? in a channel goes to Nymbot, and it answers in the channel where everyone can see it. You can also write @Nymbot in an ordinary message, or quote-reply to one of its answers to keep the conversation going — it carries up to six messages of the reply chain as context.

It is context-aware: ?ask and ?summarize receive the channel's recent messages and its active user list, so it can answer questions about the conversation that is actually happening rather than in the abstract.

The full list of commands is in the command reference.

The private chat

You can also have a one-to-one, end-to-end encrypted conversation with Nymbot, exactly like a private message with a person. Nothing you say there is visible in any channel.

Private replies are a paid feature. ?help inside that chat is free, handled on your own device, and explains the tiers, the pricing and the setup.

Credits

Two separate balances, both bought with Bitcoin over Lightning:

• ?balance — both balances. They are also shown in the chat header.
• ?buy — opens the purchase flow, with a Standard/Pro switch.
• ?gift @nym — gives credits to someone else.
• ?transfer @nym — moves your whole balance, standard and Pro, to another nym.

Pro models

?model <name> pins every reply to one specific model; ?model off returns you to standard routing. ?model on its own lists them with their prices.

Available: Claude Fable 5, Claude Opus 5, Claude Sonnet 5, Claude Haiku 4.5, GPT-5.6 Sol, GPT-5.4 mini, Gemini 3.1 Pro, Gemini 3.6 Flash, Grok 4.6, Kimi K3, Qwen 3.5 and MiniMax M3.

A Pro reply costs a per-model base of one or two credits and then scales with the length of the reply, up to a per-model maximum. The maximum is reserved when you send the message and only the actual cost is charged, so a short answer costs the base even though a long one was budgeted for.

When a model shows its chain of thought — the standard tier's reasoning route, or a Pro model that exposes it — the reply carries a collapsed 💭 Reasoning section above it that you can tap to read.

You can send Nymbot a picture, too. Every Claude, GPT, Gemini, Grok and Kimi Pro model can see images, as can the creative and translation routes on standard, and will describe one or answer questions about it.

Images and speech

Both are charged per generation rather than by reply length, and nothing is charged if a generation fails.

With a Pro model selected you can also choose the generator: ?image --model <name> <description>, picking from Nano Banana Pro, Nano Banana 2, Imagen 4, FLUX 2 Max, FLUX 2 Pro, Seedream 5 Pro, GPT Image 2, Grok Imagine and Recraft v4 Pro. ?image models lists them with prices and costs nothing.

Working in a git repository

Pro can work inside one of your repositories, the way a coding agent does. ?git connects a provider — GitHub, GitLab, or Gitea/Forgejo, which covers Codeberg and self-hosted instances — taking a scoped personal access token, a repository and a branch.

After that, messages in the repo context run as a small agent: the model lists, reads and searches your actual files. With ?git writes on it can also commit files, create branches and open pull or merge requests.

A repo task uses up to six model calls per message, each billed at the selected model's Pro price. Only calls actually used are charged.