RamelliPreviously AtlasLLMIndependent AI workspace
One space.
Every model.
Go from a quick question to a deep investigation. Thirty-plus models, a shared conversation, and a research engine that shows its work.
Keep the context.01 — 04
Standard → Deep Research
Built for
curiosity.
Ramelli puts thirty-plus models from nine providers behind one conversation, routed through a single OpenRouter abstraction. Switch models mid-thread without losing context.
It is local-first by default; your data lives on-device, with optional real-time Convex sync across devices. Four chat modes, from a quick answer to multi-step Deep Research, all share the same streaming, reasoning-visible interface.
One router, four agentic flows.
A custom, type-safe workflow engine (@repo/orchestrator) runs each request as a graph of tasks, with dependencies, retries, timeouts and parallel fan-out. The router dispatches to one of four modes; Deep Research plans searches, runs them in parallel, and loops the reflector back to the planner until the answer is complete.
- 01router
Entry point: routes each message to the right chat mode.
- 02refine-query
Analyses and clarifies the research question.
- 03planner
Plans strategy and generates search queries.
- 04web-search
Runs searches in parallel and processes content.
- 05reflector
Evaluates progress; loops back if gaps remain.
- 06analysis
Synthesises findings before the report is written.
- 07writer
Writes the comprehensive, cited report.
- 08suggestions
Proposes follow-up questions to keep exploring.
Calm on every screen.
Reasoning and web-search steps, inline sources and citations: the full research trail behind every answer.
Deep Research, fully responsive: steps, sources and the streamed report on a phone.
Nine providers, one API.
9 providers · 30+ models · 4 chat modes · 11 workflow tasks