ci/woodpecker/pr/codex Pipeline was successful
ci/woodpecker/pr/tool-rendering Pipeline was successful
ci/woodpecker/pr/web-fetch Pipeline was successful
ci/woodpecker/pr/web-search-codex Pipeline was successful
ci/woodpecker/push/codex Pipeline was successful
ci/woodpecker/push/tool-rendering Pipeline was successful
ci/woodpecker/push/web-fetch Pipeline was successful
ci/woodpecker/push/web-search-codex Pipeline was successful
ci/woodpecker/tag/codex Pipeline was successful
ci/woodpecker/tag/tool-rendering Pipeline was successful
ci/woodpecker/tag/web-fetch Pipeline was successful
ci/woodpecker/tag/web-search-codex Pipeline was successful
codex
context controls, account usage, and fast-mode requests using Pi's existing Codex login. verified extended-context profiles currently cover GPT-6 Sol, GPT-6.1 Sol, and GPT-6 Astra; unknown models are not guessed.
why
keep more context before compaction when you need it. this changes Pi's local compaction window, not server capacity or account limits. larger context may cost more.
install
pi install npm:@eleith-pi/codex
use
/eleith:codexopens the local context and fast-mode inspector, including compaction and cache status. Barecontextandfastopen the same overview./eleith:codex context onselects the verified local window when larger./eleith:codex context offreturns to the catalog baseline, confirming reductions./eleith:codex context toggleswitches the saved context preference./eleith:codex fast onchecks catalog eligibility and requests priority processing after confirmation of higher allowance use./eleith:codex fast offrequests standard processing./eleith:codex fast toggleswitches the saved fast preference./eleith:codex usageopens account allowance and reset information in the same inspector, showing a spinner while fetching.
config
set a command prefix in ~/.pi/agent/extensions/eleith.json:
{ "commandPrefix": "personal" }
then use /personal:codex.