opencode-dynamic-context-pruning
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
3recommend
0warn against
Mar 1, 2026first mention
Sep 30, 2026latest mention
Why people recommend it
“So i've taken https://github.com/Tarquinen/opencode-dynamic-context-pruning and started improving it.”
“this and openviking appear to be all I need to get a proper 'do anything' setup.”
Show 3 more
“It works great for long-horizon tasks, and feels like it saves a boatload of tokens.”
“I manage it with dynamic context pruning”
“It allows the model to remove context pollution.”
Caveats and dissent
No one in these comments warned against it.
Often compared with
Who recommends it in their own words
Threads where it came up
- You said no MCP 1 mention
- Notes on gotchas while migrating 35kb preprompts from Opus to self-hosted Ollama 1 mention
- Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k 1 mention
- Running local models is good now 1 mention
- Cache hit rates of Inference are more meaningful than the headline costs 1 mention
- GLM-5.1: Towards Long-Horizon Tasks 1 mention
Also written as Dynamic Context Pruning. Data: opencode-dynamic-context-pruning.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 7
“So i've taken https://github.com/Tarquinen/opencode-dynamic-context-pruning and started improving it.”
“this and openviking appear to be all I need to get a proper 'do anything' setup.”
“It works great for long-horizon tasks, and feels like it saves a boatload of tokens.”
“I manage it with dynamic context pruning”
“It allows the model to remove context pollution.”
“I've done some very long sessions on OpenCode with Dynamic Context Pruning. Highly recommend it.”
“It seems more aggressive.”