The always-on personal agent for strategy.
Joined April 2026
- Tweets31
- Following8
- Followers67
- Likes19
Pinned Tweet
Meet your COORDINATOR, the always-on agent for strategy.
Not built for coding or endless chats. Your COORDINATOR is built for action.
Pairing frontier LLMs with @TypeSafeAI's Jev models for rapid actions and sharper decisions, all running in the background while you work.
> give COORDINATOR your fantasy football team
> tell it to find the most optimal starting line up, sit/start values, trade solutions and more
> win your league
> repeat
Real-world decision making requires fast decision models from TypeSafe, OpenAI, Cloudflare and more!
All available in Fast Mode on COORDINATOR.
- gave @getcoordinator a mock playbook
- it planned out what to do & switched to fast mode
- fed it live plays as they happened
- decided plays in under 1 second
- noted what worked vs didn't
- i scored a fake touchdown against SDSU
Live play-calling in COORDINATOR.
COORDINATOR retweeted
- gave @getcoordinator a mock playbook
- it planned out what to do & switched to fast mode
- fed it live plays as they happened
- decided plays in under 1 second
- noted what worked vs didn't
- i scored a fake touchdown against SDSU
Live play-calling in COORDINATOR.
Fast 🏃
Accurate 🎯
No reasoning 🙅
No wasted tokens 🤯
Real world strategic AI doesn’t need language, it needs rapid decisions you can trust.
Thanks to @OpenRouter, COORDINATOR allows us to seamlessly swap between all of the latest Decision models to determine the best course of action for your next step!
No LLMs needed. Pure decisions.
COORDINATOR retweeted
Sub-second, near-real-time strategy models are here.
COORDINATOR retweeted
Instant play-calling in COORDINATOR is here:
Start a mission -> Tell it what to decide -> Done in milliseconds.
All rooted in whatever data and context you give it.
Try it 👇
Decisions, made instantly. Live today.
Bring your context, set your preferences, and let COORDINATOR make the calls with you –– fast.
Start now at coordinator.co
58
Context is what makes it your call.
Missions pull your playbooks, cheat sheets and scouting reports straight from your Library, then layer in memory: run-heavy, no QB runs, go for it on 4th & short.
Every quick-call is grounded in your files and your rules.
1
86
Switch a mission to fast mode and every message routes to a dedicated decision model, not a slow LLM.
The strategy math runs on deterministic tools, meaning you won't get hallucinated answers.
Oh, and it's faster than a play clock.
1
141
Make decisions instantly with COORDINATOR.
COORDINATOR runs the latest decision models from OpenAI, TypeSafe’s Jev, Cloudflare’s Clef and more, so you get live, split-second calls built on your context and your preferences.
Learn more below 👇
COORDINATOR retweeted
@getcoordinator is sick. Just asked it to optimize my Fantasy lineup for week 5. Only had to give it a link to my team and league.
docs.google.com/document/d/1…
You can experience the power of Jev's decision making for free inside COORDINATOR by going to:
coordinator.co
15
Jev decides if COORDINATOR should skip replying to simple messages entirely.
Meaning will not get a token-wasting reply from a simple, "thanks!".
This is because Jev checks every turn to determine if the message is important enough to reply to or not.
1
18
We've found Jev makes more accurate decisions for COORDINATOR compared to traditional rules and small classifiers.
Meaning COORDINATOR users can feel more confident that their agent has made the right call.
1
12
To start, every COORDINATOR mission goes to the right worker.
Jev decides.
Code executes the work.
An LLM writes the answers.
A frontier LLM doing all three is incredibly expensive, slow and can provide wrong answers frequently.
1
74
Most of what our agent does isn't writing, it's rapidly deciding.
"Is this worth a replying to? 🤔"
"Does this mission need tools? 🛠️"
"Is this response actually the answer? 🤷"
Jev makes COORDINATOR decide faster, while saving money in the process.
Here are some examples 👇
COORDINATOR retweeted
- 20,000 rows & 30 columns of mock scouting data for hundreds of athletes
- asked to look for every player rated 3 & 4 stars, w./ a 4.70 yard dash or faster, in Texas, Arizona and California
- verified by Jev in 1 second
- report generated in 1 minute
@getcoordinator
COORDINATOR features a list of Jev sub-agents to rapidly and accurately perform various tasks like sorting through data, triaging and judgement, verifying answers, and more.
Each one can perform operations in milliseconds compared to LLMs
Create your personal always-on strategy agent now.
Use code COORDINATOR to get one month free of all the frontier models for your various strategic needs.
coordinator.co
1
37