CAPABILITIES · SPEC SHEET

Everything an instance
can be told to do.

Each capability is a config seam: off by default, generic, and yours to turn on. Nothing below is hardcoded to any deployment.

01

Chat & output

The core experience — and it is genuinely nice to use.

ACTIVITY TRACE

Watch it think

Answers stream token by token. Every tool call surfaces as a labelled chip; runs of read/search steps collapse into a single "gathered context" row so long investigations stay scannable.

  • Live elapsed timer and status
  • Colored, readable Mermaid and wrapping tables
  • Per-turn cost + copy-answer
activity trace
Live activity trace with tool chips
THEME

Themes, done properly

A semantic-token palette with dark, light, and system modes — applied before first paint, no white flash. Admins get a Theme dashboard: templates, code themes, and markdown accents.

Settings · Theme
Admin theme dashboard
02

Acting safely

Let the agent do things — on your terms.

approval
Interactive approval card
APPROVALS

Interactive approvals

List a tool under approval_required and the turn pauses with a card: the exact command, and Approve & run / Deny — decided by the thread owner or an admin. No decision in five minutes denies; a page refresh keeps the card decidable.

  • Read-only mode always wins
  • Survives refresh, replayed into the trace
03

Reaching out

Channels, embedding, and attribution.

Channels

GitLab & GitHub

@mention the bot in an MR/PR or issue comment; it replies inline with one-conversation-per-target memory. HMAC-verified, allowlisted, Enterprise-ready.

Embed

Embedded mode

Load with ?embed or in an iframe and the chrome yields to the host portal — drop the chat into any app you already have.

Usage

Per-origin cost

Every turn is attributed to its surface — web, embed, a channel, or a custom label. The admin usage table grows per-origin cost columns automatically.

04

Working across repos

Turn the agent into an operator — with a toolbox and a leash.

WORKSPACE

Multi-repo workspace

Declare a set of GitLab/GitHub repos and the agent keeps them checked out — investigate, branch, commit, and (per the instance's autonomy setting) push and open MRs/PRs.

  • Gated autonomy pauses pushes and MR creation on an approval card
  • Pushing a default branch is refused in every mode
  • Ephemeral cache — remotes stay the source of truth
TOOLBOX

CLIs by declaration

Instances declare the CLIs the agent may use. Catalog entries like kubectl@…, helm@…, terraform@… are fetched once with vendor-checksum verification; custom binaries by url + sha256.

  • Declared tools lead the agent's PATH
  • An allowlist decides what runs — available ≠ permitted
  • agentkit discipline hooks keep autonomous git clean
05

Model backends

Bring your own model access.

Anthropic

Claude, direct

The default path — an OAuth token or API key and you are running.

Bedrock

AWS-hosted Claude

Offer Bedrock inference profiles in the picker; auth from the pod's AWS credentials.

Roadmap

More engines

An engine seam keeps the door open to additional backends behind one interface.

Deploy neutron →