>jevable

Make your monitor smart.

Your agent pays a full turn for every line that wakes it. jevable reads each event by meaning, for a fraction of a cent, and wakes the agent only when it matters.

Paste this into your agent
Set up jevable: run `npx -y jevable guide` and follow it.

Works in Claude Code, Codex, OpenClaw, Hermes, pi, dsh — any agent that runs shell commands. It asks you once for an API key from TypeSafe, OpenRouter or Vercel, and sets itself up.

7.3× fewer agent wakes. 85% recall. View benchmark →
01What a wake costs

Every wake is an agent turn.
Pay only for the ones that matter.

Woken, an agent reads its context and thinks — cents to dollars a turn. A Jev question reads one event for about $0.000023. With jevable in front, the agent watches everything and spends a turn only when something counts. See it measured on real events →

  • Wake on every event1,000 agent turns a dayat once, and for everything
    $3,000a month
  • Check every 30 minutes48 agent turns a dayup to 30 minutes late
    $144a month
  • jevable in front10 agent turns a daywithin a second, only what matters
    −98%$30.68a month

jevable is charged a Jev question for every event here; plain conditions in the rule usually spare most of them. One question cost $0.000023 on average in the demo below. Many matches a day? --window or --cooldown fold them into fewer wakes.

02Same events, two alerts

grep wakes you for words.
jevable wakes you for what they mean.

A keyword alert fires on a teaser and a pun, and stays silent when an outage is worded differently. The same events through a jevable rule:

“Tell me when Tibo announces that Codex usage limits were reset.”

16 real events · @thsottiaux on X
eventgrep -iE 'reset'jevable ≥ 0.7

  1. This was fixed. You know what's coming 👀 Give us 24 hours to reset the Codex rate limits across all plans.

    woke · noise0.00

  2. Hi. Over the last 24 hours we had three separate small incidents that affected Codex reliability. Those are three too many and we are taking active steps for them to not reproduce. I have reset usage limits for Codex across all paid plans. May the tokens flow again.

    woke1.00

  3. As we are still investigating, I have reset everyone's Codex usage limits. This is a hard reset given some users had stacked up to three banked resets already that they can apply on their own schedule.

    woke1.00

  4. Reset your expectations: the new model is much faster than you think.

    woke · noise0.00

  5. Codex now supports plugins in the CLI. Try it with codex plugins and tell us what you build.

    0.00

  6. Dearest gentle codexer. We did a sneaky double reset. Not only do you get a full reset on us. But you are also getting one into the reset bank to use at your own leisure. Enjoy

    woke0.97

  7. Hi! To celebrate its 1-year anniversary, I have allowed Codex to reset its own rate limits across all plans. Enjoy all the new features.

    woke0.98

  8. Old news actually from a bunch of days ago, but crossed that 15M. Enjoy a nice reset everyone. Landing in the next hour or so, go /fast.

    woke0.87

  9. Another reset for our Codex and ChatGPT Work users. Actually hit 9M active users way earlier today, but then got distracted by the approximately millions of things the team is doing to keep the systems up and reliable. Should have that sweet 100% weekly usage limit back in a few minutes.

    woke1.00

  10. We are investigating elevated error rates in Codex. Updates to follow.

    0.00

  11. Hello. We have reached 8M active users across Codex and ChatGPT Work. We are once again resetting the usage limits for all. And we continue to not have the 5h rate limit as well, allowing everyone to explore the boundaries of GPT-5.6 Sol and discover how ambitious you can be.

    woke0.99

  12. I previously promised a reset for every 1M in additional active users for Codex, until 10M. We blew past that and have been silent since 10M. Little surprise for you tomorrow.

    woke · noise0.04

  13. What should we build next for Codex? Replies open.

    0.00

  14. Oops... I did it again. Enjoy reset usage limits for all paid users for Codex and ChatGPT Work. Super grateful for an incredible team who is iterating at lightspeed and keeping the infra up as we scale faster than ever. Enjoy the weekend!

    woke0.97

  15. People keep asking about usage limits. We hear you, and we are working on a lot more capacity.

    0.00

  16. We are reseting usage for all paid users of Codex and ChatGPT Work. Please continue reading for an update on Codex usage limits. The team has been working around the clock, going through thousands of reports and shipping fixes. Depending on how you use Codex, you should see your usage go between 10% and 50% further than before. We really went with a fine comb, with many uncovered small things being longstanding and here is what we found and fixed: - Compaction. We were keeping old images during compaction, sometimes making the context large enough to trigger compaction again. After the fix, usage dropped around 10% for users making heavy use of images. Fixed. - Memory. Background memory workers could inherit Stop hooks and keep running when the hook wouldn’t let them finish. This affected fewer than 1% of users, with the long tail being pretty bad and we saw one example thread check whether it could stop 15,000 times. Fixed. - Goals. In some cases, a set /goal could finish and then keep going past the intended stop condition, or the model would keep retrying broken tools without stopping. We saw examples consume anywhere from 15% to 70% of a weekly allowance. Fixed. - Automations. Some custom schedules could run more frequently than configured. Fixed. - Subagents. Smaller models (e.g. Luna) sometimes picked more capable helpers without being explicitly asked. The same was true where the orchestrating model not running in /fast mode could request sub-agents to run /fast. Fixed. - Computer History. The older implementation could lead to repeatedly summarizing overlapping activity. For some cases we saw it consume up to one fifth of the weekly usage per week. Fixed. - Rolling task summaries. Ordinary turns were triggering extra background requests. These added about 1% to token usage. Small each time, but it adds up. We have disabled this. - MCP. Some tool results could be encoded twice. We also found tool instructions getting cut off and fetched again. Fixed. We’ve also made architectural changes to prevent these from regressing and our teams will get paged if it happens regardless. We are also working on showing you directly in the app where your usage goes so you don’t have to guess. Goes without saying that we’re resetting usage limits and I hope you enjoy a very nice Saturday!

    woke1.00
grepwoke you 12×3 for nothing
jevablewoke you 9×all real, none missed16 questions to Jev, $0.00030 in all
The rule
judge.choice(line,
  "Does this post announce that Codex usage limits are being reset?",
  {"reset_now": "The limits have been reset, are being reset right now, or will be back within the hour (e.g. 'back in a few minutes')",
   "reset_soon": "A reset is promised or teased for later (hours or days away) but has not happened yet",
   "other": "Anything else, including complaints about limits or other uses of the word reset"})["reset_now"] >= 0.7

Real posts, incidents and issues, labelled before the first run. Scores are real Jev answers, recorded when this page was built. “marker” means a plain condition in the rule decided without asking Jev.

03How it works

It is a pipe: one question.

  1. Source

    Anything that prints one event per line: tail -F a log, an API polled in a loop, a feed.

  2. Question

    Like a grep pattern, but a yes/no question that Jev answers for each line. Do the exact part with grep or jq first.

  3. Wake

    Each match reaches the agent the way its runtime listens — or goes straight to your phone.

Live run“Wake me when a Hacker News comment reports a problem with Claude Code.”
  1. 100comments came inthe source
  2. 41mention Claude Codeplain condition, no Jev call for the other 59
  3. 10report a problem with itone question to Jev each

10 wakes, each one real.

# wake when a log line reports an outage
tail -F app.log | jevable "Does this report that a dependency is down?"

# JSON: judge one field (a jq path)
gh api repos/o/r/pulls/12/comments --jq '.[] | @json' \
  | jevable --on .body "Does this ask for a change to the code?"

# -v inverts, -t sets the threshold, like grep flags
git log --oneline | jevable -t 0.8 "Does this commit change a public API?"

# more than one question: a CEL rule
jevable filter 'json.status == "resolved" && judge.boolean(json.name, "…") >= 0.7'
~0.3 s
per question to Jev
$0.000023
per question, measured
0 calls
when a plain condition decides
your fields
only what you point it at is sent
04Before you trust it

Know the threshold, not a guess.

The agent writes samples that should and should not wake you, then runs jevable test. It shows every score and the thresholds that separate the two sides — here, anything between 0.03 and 0.89.

$ jevable test -f rule.cel --yes yes.txt --no no.txt
ok  yes  reset_now  1.00  Another reset for our Codex…
ok  yes  reset_now  0.89  Old news actually from a bunch…
ok  yes  reset_now  1.00  We are reseting usage for all…
ok  no   reset_soon 1.00  You know what's coming 👀 Give us…
ok  no   other      1.00  Reset your expectations: the new…

16 of 16 as expected.

["reset_now"] yes 0.89–1.00 · no 0.00–0.03
  → any threshold above 0.03 and up to 0.89 separates them
05Runtimes

Wakes the agent you already use.

The guide tells each agent how its own runtime gets the events back — while the session is open, or for days after it closes.

06$JEVABLE

A wake costs money.
The token pays for the watching.

Every wake jevable sends is one question to Jev, and every question has a price — about $0.000023, measured on real events. $JEVABLE is the prepaid credit for that work: the fuel for monitors that stay awake while your agent does not.

$0.000023
cost of one judgement
24/7
watches that never sleep
Solana
chain — launching soon
SPL
token standard
Contract address · Solanacoming soon
$JEVABLE launches on Solana. The CA is posted here and on @jevable the moment it goes live.

Nothing here is live yet: no tier is purchasable, no price is set, and no $JEVABLE contract exists. If any line reads like an investment, the line is wrong. Follow @jevable →

One line in,
a smart monitor out.

Describe what to watch. Your agent writes the rule, tests it and arms it. It asks you once for an API key from TypeSafe, OpenRouter or Vercel, and sets it up.

Paste this into your agent
Set up jevable: run `npx -y jevable guide` and follow it.

Works in Claude Code, Codex, OpenClaw, Hermes, pi, dsh — any agent that runs shell commands. It asks you once for an API key from TypeSafe, OpenRouter or Vercel, and sets itself up.