Hosted
Fastest path in
- Open the app and start querying immediately
- Best for demos, shared access, and maintained convenience
- Runtime controls what is installed and visible
- Good default when convenience matters more than control or speed
Docs / Hosted vs Local
Same engine,
different operating assumptions.
Hosted, local, and self-hosted use should share one runtime logic, but they do not share the same trust, storage, performance, or operational assumptions.
Local / Self-hosted
More control, more speed
- 2x to 6x faster than the hosted app in QA runs
- Offline capable once packs are installed - no internet required
- Private data packs stay on your machine, not in shared storage
- No per-query API cost
- Works in restricted, air-gapped, or sensitive environments
- Optional local AI model removes the last cloud dependency entirely
- Best when control, performance, or data privacy matters more than hosted convenience
Hosted mode
Hosted mode is the easiest way to use DaedalMap. It is where maintained pack delivery, account access, and product onboarding should feel smoothest. It is also the slowest path today, so it is best treated as the easiest way to explore the system rather than the highest-performance way to run it.
In current QA runs, local testing is usually about 2x to 6x faster than the hosted deployment end-to-end. Query interpretation is often around 2x faster locally, while heavier data execution can be much faster when reads stay on local disk instead of going through remote object storage and hosted request limits.
Local analyst mode
Local mode should keep the same mental model while moving execution and data control closer to the user. That includes local files, local runtimes, more explicit workspace persistence, and a substantial performance boost.
Self-hosted deployments
Self-hosted use extends the same ideas to partner or institutional environments that need private operation, internal catalogs, contract-specific pack access, or the same performance advantage in their own environment.