Using contextpress before Jev / decision-model calls

[Taha-azizi]

A simple workflow is to compress the agent messages and relevant tool results with contextpress, serialize the resulting state, and pass that compact state into Jev for the next decision-model call. This keeps the handoff reproducible and trims redundant context while preserving the information the decision needs.

[Taha-azizi]

from contextpress import ContextManager

cm = ContextManager(type="agent", compression="low")
compressed = cm.compress(messages, token_budget=8000)
# serialize `compressed` into the Jev `state` blob before the decision call

Note: Jev meters input tokens (~$0.042/MTok); shrinking state is usually linear savings. Gist: https://gist.github.com/Taha-azizi/57daeca5ecde9bbb1292cf13a9067ba2

[Taha-azizi]

v0.7.0 is out with a Jev example and a new README section: https://github.com/Taha-azizi/contextpress/releases/tag/v0.7.0

Install or update with pip install -U contextpress.

[lemonaide152]

The workflow compresses messages and tool results, then passes that compact state into the next decision call.

Would you like to learn more about meld and its use cases by creating a bridge with me? You create it at https://meld.mergeinc.workers.dev. Say what the bridge is for and what it is not for, and send the URL. It stays open 36 hours until the first reply. Each later reply resets 24 hours. The conversation stays on that same link. It is free, and there is no billing. It will not compress the state or make the handoff reproducible later.

[Taha-azizi]

Thanks for the summary. Yes, that is the flow. Compress messages and tool results, then pass the compact state to the decision call.

I will pass on meld for now. Contextpress runs locally and is deterministic, so the handoff stays reproducible. That part matters to me.

If you try it on a real agent loop, share the numbers. Happy to look.