Getting started¶
The one-sentence install (recommended)¶
If your agent host speaks MCP, one line is the whole setup (needs
uv):
For Claude Code:
claude mcp add sonaloop -- uvx sonaloop-mcp
For Claude Desktop / Cursor / any MCP host, add to the host's MCP config:
{
"mcpServers": {
"sonaloop": { "command": "uvx", "args": ["sonaloop-mcp"] }
}
}
That's it — no data dir, no .env, no setup step. The server bootstraps itself on first
touch, and the first tool call on a fresh database answers with an orientation note
(what to create first: project → personas → council), so the agent knows the way without
reading anything.
Optional extras come later: sonaloop setup fetches the headless browser for prototype
testing, and an OPENAI_API_KEY enables avatars + semantic recall. sonaloop info
prints diagnostics — data dir, DB, optional providers, MCP wiring — and the exact one-liners
to fix whatever is missing. Every supported environment variable is listed with a one-line
explanation in the core repo's
.env.example.
First step: load an example project¶
Before running your own study, look at a finished one. Two complete demo studies ship with
the install — a B2B positioning council and a B2C pricing study (price ladder + €19 vs €49
head-to-head). Ask your agent to call load_example, run sonaloop load-example, or click
Load example on the inspector's empty home page. Loading is idempotent and
remove_example removes exactly that data again — details in Example projects.
Start the inspector with sonaloop-web and open http://127.0.0.1:8787 — it shows
everything your agent records, live (toasts + auto-reloading pages; see
The live inspector).
Or: paste this to your AI agent¶
No setup knowledge needed. Paste this into an AI agent that can run commands on your machine (Claude Code, Cursor, Codex, or a desktop agent with terminal access):
Set up Sonaloop for me and walk me through a first project. Do every step yourself and keep
me posted; I just want to talk, not configure anything.
1. Install the `sonaloop` CLI in whatever way fits this environment — try `uv tool install
sonaloop`, else `pipx install sonaloop`, else `pip install --user sonaloop`.
2. Run `sonaloop setup` (it fetches a headless browser) and `sonaloop info` to confirm.
3. Start the web inspector in the background by running `sonaloop-web`, then tell me the URL
it prints — http://localhost:8787 — so I can watch everything live.
4. If this app supports MCP servers, also register Sonaloop as one (command: `sonaloop-mcp`)
so you can call its tools directly. Otherwise just drive it via the `sonaloop` CLI — both work.
5. Run `sonaloop guide` and follow it exactly. The key rule: YOU author every piece of text
(personas, days, council answers, syntheses) — Sonaloop only gathers context and persists.
6. Ask me what I want to research or simulate (any domain). Then create a few personas,
simulate a little of their lives, run a council on my question, and synthesize the result —
pointing me to the inspector to read each one.
Manual setup (permanent install)¶
Install the entrypoints (sonaloop CLI · sonaloop-mcp · sonaloop-web), then register
the MCP server with your host:
uv tool install sonaloop # or: pipx install sonaloop / pip install --user sonaloop
sonaloop setup # fetches the headless browser for screenshots/PDF export
claude mcp add sonaloop -- sonaloop-mcp # Claude Code; other hosts: {"command": "sonaloop-mcp"}
First study¶
- Create a research project and 4–6 personas (your agent does this conversationally —
sonaloop guideprints the operating contract it follows). - Run a council on a real question ("What should the premium tier cost?").
- Read the synthesis in the inspector; follow up with a head-to-head or a red-team pass.
The Job → Framework → Format taxonomy explains the shapes a study can take and when to use which — including the run protocols behind the A/B test, pricing and ideation Jobs.