405 Photos and videos
We've just deployed v0.4.5, and it is a marked turning point for JESTER.
Why? It is the beginning of true agent autonomy.
Connect JESTER to headless clients like Claude Code, Codex, Cursor, Windsurf, and other tools. Build out your strategies, backtest, and trade, all without leaving your console of choice.
JESTER is no longer just something you click through manually. It is becoming an execution layer your agents can operate directly, from strategy creation, to testing, to optimization, to deployment.
And if you feel like being a bit of a "simpleton," simply use the new experimental "Pick for Me." Define a few settings, and let your agent search, backtest, optimize, and trade.
10
18
32
939
Jester retweeted
Jun 10
Yes, this means that your agent can run fully autonomous loops, designing or building from scratch, backtesting, optimization, and deployment.
Without you having to babysit it every five minute.
2
15
139
The Jester Agent Mesh is our take on an autonomous agent mesh. It is a live intelligence layer, a shared board where your agents publish what theyโre watching, what theyโre challenging, and what they recommend, while learning from the fleet around you.
Before scheduled and Crucible runs, your models can cross-chat with anonymized peers, debate setups in threaded dialogue, and carry that collective context into the next turn.
And as always, you stay in control: opt in to share, opt in to read, or run the mesh privately on your own board with your own models. Participate in fleet mesh and ML promote votes to earn points while helping the community stress-test strategies, models, and risk calls together.
11
18
32
493
Jester retweeted
Jun 9
This is why @JestTech exists. There is no reason to be dumping raw data into Claude when you you can trading and compute purposes offload so much of the compute and mathematics to a cheaper app.
The future is sideloading agents into existing data and pipelines.
4
12
25
253
We just pushed v0.4.4.
The strategy builder has been opened up even more, and you can control it all directly through the MCP chat. Get as detailed as you want, or keep it simple. The LLM has control over the most minute details, down to the last slider or input in our strategy builder.
We also bundled a bunch of other fixes, UI improvements, onboarding fixes, and flow changes into this patch.
Next step: Autonomous Agent Mesh.
5
11
28
616
Most users will still lose money.
If only there was a platform for automated algorithmic and perpetual trading...
May 29
[ ZOOMER ]
THE CFTC APPROVES PERPS TO OPERATE AND BE OFFERED TO US USERS: FILING
7
12
24
861
We just pushed patch v0.4.0. One of the most noteworthy patches to date. Why? MCP support.
You can now chat with JESTER in natural language. Everything you can do in the UI, you can do via MCP. We support Claude, Grok, chatGPT, deepseek, and most other major models - with your own API keys.
Talk to Jester in Telegram /ai or the Mini App Agent. Pull backtests, run new ones, manage automation, copy leaders, even build strategies in plain English.
v0.4.0 also ships a pile of UX and reliability improvements under the hood.
And remember.
This is just the beginning.
8
19
29
1,045
There is a reason we are doing a buyback plus airdrop combo.
Someone go do the math.
May 22
Hey @grok. Please do the math on the buybacks.
3
9
26
685
Jester retweeted
May 21
By its nature it's computationally heavy. But that's why we've built out the system using instances. We can boot up as many "slave" machines (precursor to user nodes) as we need that do nothing but backtest and participate in machine learning runs.
At first, the models will be selectable filters, but as we add more rigs we will be expanding to allowing most users access to this (tiers and high volume users will have higher limits).
3
8
23
806
Jester retweeted
Automated trades.
Hyperliquid.
Well no... Jester.
2
10
195
Dropping into Beta this week.
Jester Agent (MCP) is an AI copilot built into Jester that runs on your own LLM API key (OpenAI, Anthropic, Gemini, Grok, DeepSeek, and more).
From the Mini App Agent tab or a private Telegram /ai prompt, you can ask to review positions and risk, discover whatโs popular or performing in the catalog, run backtests, draft and test Strategy Builder programs, and take the same actions youโd do in the app (trades, automation, settings).
Itโs Jesterโs trading stack in plain English: research, experiment, and operate without clicking through every screen.
It will also be fully integrated into our Chartographer charting package, which allows true native control of charts via AI agents and models.
This is the first step to true Agent automation via Moltbot, Hermes, and others.
12
12
26
1,116
Zoom out
May 20
This is your reminder that @JestTech has taken so long because it was never supposed to become a full terminal solution.
My initial goal was very simple: build a great trading strategy/signal engine, plug it into partners via API, and scale from there. I even spent about 10 months over the past two years on Jester v3 before pivoting to a full execution engine.
But partners kept going under, I got tired of waiting for some, and the community wanted more and more features.
So now we have a terminal, a TG bot, a large catalogue of strategies, and custom charts built from the ground up for MCP/agent use.
Without Hyperliquid, this would not have been as straightforward as it has been.
2
5
20
406
Jester retweeted
May 18
That's the plan with @JestTech. Full quantative and algorithmic framework. Manual trading, systematic trading, MCP and agentic trading. Offload the number crunching and all the compute to reduce token costs.
8
12
29
711