simulated workspace — scripted world, real productsimulated demonew here? take the walkfor agents · connect your AI
architect

workspace settings

Personality + team type compose the dialect every surface renders under. Personality tunes the recall lens (weights, tag + kind affinities); team type drives parser heuristics during import and priority widget defaults. Changes take effect on the next request.

Personality

The workspace lens. Applies to every tenant unless they set a personal override on /settings/me.

Team type

Drives parser heuristics during import, seed-template choice, and priority widgets.

Managed AI — You're using eleven11-managed AI (shared Dhara Ollama). Usage is metered against your eleven11 invoice. Switch to your own provider key any time below — your existing memory keeps its embedding.

LLM provider

Where every workspace `:ask` dispatches. dhara-ollama is the free sovereign tier; the three BYOK options expect your own API key. Per-canvas overrides on `/c/<slug>` take precedence.

Current dhara-ollama · (provider default) · no key needed

dhara-ollama is shared infra — no key needed. Saving will clear any stored BYOK key.

Memory embeddings

Embedding provider for semantic memory recall. Historical rows stay embedded under the model they were written with — switching providers narrows vector recall to new rows until a re-embed pass runs.

Current (not configured) · (not configured) · not configured
Model
nomic-embed-text
Model selection is fixed in this release.

Dhara-managed uses the operator's shared endpoint — no key needed. Saving will clear any stored BYOK key.

Notion

Connect a Notion integration token (PAT). Architect can then sync accessible top-level pages + databases into this workspace via the import wizard. Tokens are encrypted at rest and never leave the server.

Create one in Notion: Settings → Integrations → My integrations → New internal integration. Then on each Notion page or database, Share → Add the integration so it can read it.

Current Not connected