AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
[ ] Organize the application state into well-defined components
Update: the easiest way to realize this seems to be autogpt.core (#4770)
[ ] Implement a persistence mechanism
[ ] Implement a rehydration mechanism
[ ] Allow application state to be exported/transferred
[ ] Allow resuming a run from any point in the execution history (important for debugging/development)
Note
This issue has been converted to a roadmap item by a maintainer (@Pwuts). Thanks for creating this overview @Boostrix! <3
Original description
meta issue to keep track of everything that's been said in the context of suspending/resuming previous [crashed] agpt sessions - ideally with a capability to resume the agent so that it can validate its known workspace artefacts and use not only any data files it has used previously, but especially any custom files, and especially tools/scripts it may have generated - this implies some fort of dependency tracking (e.g. timestamp between folders/files).
Related
430
822
1176
1531
3382
3542
3933
4063
Keywords for people searching: saving, resuming, checkpoint, storing, continuing, serializing, pickle
Motivation π¦
We need to be able to deal with different forms/types of stateful behavior, for example:
the actual goal/objective at hand
LLM interpretation
local results errors/warnings/success
side effects (think file system, directories/files)
This issue has automatically been marked as stale because it has not had any activity in the last 50 days. You can unstale it by commenting or removing the label. Otherwise, this issue will be closed in 10 days.
Objectives π
autogpt.core
(#4770)Related
430
822
1176
1531
3382
3542
3933
4063
Keywords for people searching: saving, resuming, checkpoint, storing, continuing, serializing, pickle
Motivation π¦
We need to be able to deal with different forms/types of stateful behavior, for example: