Skip to main content
The Sigilix CLI and Deep-Research Chat are part of the private beta. Availability is per-account during the beta — join the private beta to get access, and contact support@sigilix.ai if you are already in and want it enabled.
The hosted GitHub App reviews every PR. But review is not the only place the depth pays off — and the PR thread is not the only place developers work. The Sigilix CLI brings the same review to your terminal, and Deep-Research Chat lets you interrogate the repository and its review history in plain language. Both draw on the earned-context layer: the index, code graph, trust ledger, review memory, and evidence manifests that every review deposits. The terminal and the chat read from that understanding instead of rebuilding it.

​
Two surfaces, one understanding

Sigilix CLI

Review parity in the terminal. Run a Sigilix review on local changes before you open a PR — the same five-specialist ensemble and believability gates, reaching for the same earned context the hosted reviews built.

Deep-Research Chat

A chat grounded in your repo and its review history. Ask why a finding fired, what a change affects, or how a subsystem fits together — answers are anchored in the index, code graph, and past reviews, not guessed.

​
Grounded in earned context

The difference between the Deep-Research Chat and a generic code chatbot is the substrate. A generic chatbot reads whatever fits in its window. Deep-Research Chat reads from the earned-context layer:

The repo, in context

The code graph and index mean answers reason about callers, dependencies, and symbols across the whole repository — the same context-before-judgment that powers reviews.

The review history

Past findings, their proof-tier receipts, and taught rules are retrievable. The chat can tell you not just what the code does, but what Sigilix has already learned about it.

​
Bring your own models

Sigilix is not locked to one provider for the CLI and chat. You can bring your own:

Codex CLI

Use OpenAI’s Codex CLI as the model behind your terminal workflow.

Claude Code

Use Anthropic’s Claude Code as the agent driving the work.

Your own SDK / keys

Bring your own SDK or provider keys and route through them directly.
BYO models apply to the CLI and chat surfaces. The hosted PR ensemble runs models tuned per specialist role — see The Ensemble — with cross-provider fallback so one provider’s outage cannot silence a specialist.

​
How it fits together

1

Reviews build the context

Every hosted review deposits a verified layer — index, code graph, trust ledger, review memory, evidence manifests.
2

The CLI reviews locally

Run a review from the terminal before opening a PR, reaching for the same earned context, using the model you bring.
3

The chat answers from history

Ask the Deep-Research Chat about the repo or a past finding; it answers grounded in that same layer.

Earned Context

The reusable layer every review deposits — and what the CLI and chat read from.

The Ensemble

The five-specialist architecture behind review parity.