Elaaf retweeted
One overall accuracy score can hide serious document errors.
MiniCPM-OCR scores equations and tables separately from plain text, using individual 0–1 scores for each. That matters because a table can be badly misread even when the rest of the page looks perfect.
For production document AI, granular verification is far more useful than a single average. #AI #OCR
🎬 MiniCPM-OCR in action — running locally on iPhone Duo and iPad.
Parse fast and accurately. Guard the results. All in your hands. 🙌
Handwritten notes, complex formulas, newspapers, poor-quality scans—real documents put OCR to the test.
📄 MiniCPM-OCR 1.0 — SOTA OCR performance and speed at 1B scale.
It achieves 93.68 on Real5-OmniDocBench and 54.7 on OCRBench v2, combining accurate document parsing with efficient inference.
🛡️ MiniCPM-OCR Jev — A 2B decision model that checks OCR accuracy.
It approaches Qwen3.5-397B-A17B on JevOCRBench and works with outputs from different OCR parsers.
Watch it compare parsed text, formulas, and tables with the original image and assign 0–1 correctness scores—helping you pick better results and identify what needs human review.
🔓 Open weights. Deploy your way.
Run with llama.cpp, vLLM, SGLang, or Ollama. Customize with open-source fine-tuning code.
Watch it in action. 👇
🔗 GitHub: github.com/OpenBMB/MiniCPM-O…
🤗 Hugging Face: huggingface.co/spaces/openbm…
👉 Try the Demo：platform.modelbest.ai/consol…
Use the API：platform.modelbest.ai/consol…
Elaaf retweeted
what do you mean i can create an agent that tailors your application and applies for jobs???
firstprompt’s application prep helps you lay out your real experience clearly and shape it around the job you’re applying for, so you stop sending the same generic application everywhere
pick the next application you want to send and start free
firstprompt.build/?8x=yych4d…
🤝 Paid partnership
“just use ai”
okay bro i opened claude, now what
firstprompt gives you courses + ready-made prompts to build a website, prepare for your next job or automate work for your business
pick something you actually need done and start free
firstprompt.build
Elaaf retweeted
Hey Stella Fam👋:
I want to introduce you to a wonderful project. Are you ready to fly with it 🚀🚀
🔥 $sorl to moon
✅Main Network:
Base (Ethereum Layer 2 Blockchain)
✅Smart Contract Address: 0x653ba10378ed53bb78249bebe4c533f0d5bf515f
✅For more updates Visit X:
x.com/SORLGPATECH
Elaaf retweeted
What if your voice could handle more than just conversations?
With SpeakON, you can capture ideas, write messages, take notes, translate conversations, and automate tasks — all with one AI-powered tool.
Less effort. Fewer apps. Smarter productivity.
What feature would you try first?
#ad
🤝 Paid partnership
An idea, a client message, a team follow-up.
Explore SpeakON with Fred, Kiki, and Rich: voice typing, Notes, translation, offline recording, and Agent workflows in a working day.
What would you try first?
speakon.app/x
Elaaf retweeted
Speed makes a big difference when creating AI videos.
Pika H3 Flex is showing around 1.8× faster DiT denoising than Fal H3 Max at 768p, reaching nearly 2× speed for longer video generations.
More time creating, less time waiting for results. That’s what makes this interesting.
🤝 Paid partnership
Meet Pika H3 Flex, debuting today as the #1 open weight model on DesignArena’s t2v board.
It’s built for professional iteration, with exportable audio stems, resolutions up to 4k, and faster-than-realtime generation. It can produce a 15-second clip with soundtrack in as little as 7 seconds.
So, yeah. It’s kind of a flex 💅
Elaaf retweeted
⚡ THE PIKACHU CARD HUNT IS LIVE ⚡
One of the most iconic characters on the planet just got its own Solana memecoin.
$PIKACARD — Pikachu Card Coin
Global recognition + Pokémon card culture + pure Solana memecoin energy.
Early community. Early chart. Early entry.
One Pikachu. One community. One electric run.
CA: 39GPZUbQPmPPTAAxpqdgGcuujN1iqpqkZ9iqSb9Jpump
X: xitter.cf/Pikachucardcoin
Catch it before the rest of the timeline does.
Let’s make some noise. 🟡⚡️
#Solana #MemeCoin #Pikachu #PIKACARD
Elaaf retweeted
Not every teammate is online at the same time. A project can move from one time zone to another, but the context often gets stuck with whoever was there first. Imagine leaving the day's notes and unfinished tasks in a shared room, where another teammate's agent can pick things up before they even log in. That's a practical use case for S.Space that I can get behind.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
A useful team discussion can produce a great plan, a table of decisions and a dozen follow-up tasks. Then the chat keeps moving, and finding those things later becomes another task. I like that S.Space puts emphasis on where the work actually lives. Shared docs, pages, tables and boards give people and agents something concrete to keep building on. The conversation is important. What comes out of it matters just as much.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
One underrated problem with having multiple AI agents is that you end up managing the conversations between them. You ask one to research, another to organize, then manually carry the context over to the next. Imagine working on a project where your teammates and all your agents can simply join the same room. Less context shuffling, more actual progress. That's an interesting direction for S.Space.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
Picture a small team planning a product launch. One person is working on the messaging, another is checking the timeline, and their agents are helping with research, drafts and details. The annoying part usually isn't the work itself. It's keeping everyone aligned across different conversations. Having the people and their agents work in one shared room, with the plan staying in Commons, makes a lot of sense to me.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
The real cost of using several agents isn't always the prompts. It's the context you have to keep moving between them. Dot might know one part of a task, Muse another, and Grokbot something else. But the project itself shouldn't be fragmented just because the agents are separate. The idea of bringing them together with your teammates' agents in S.Space is worth exploring.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
I've been experimenting with S.Space for a research workflow. The idea is to have one agent collect information while another helps organize the findings. Still figuring things out, but I like where this is going.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
I tried setting up a workflow where one agent handles research and another helps turn the findings into content ideas. Having everything in one shared space feels like a much more natural way to work.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
Been playing around with the idea of connecting my coding agent and research agent in the same room. I want to see how well they can share context without me constantly stepping in.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
The voice feature is what I wanted to experiment with first. Being able to talk through a task instead of typing instructions to every agent separately could be really useful.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
I used S.Space to explore how multiple agents could help organize a project. What interests me most is having the discussions and actual work happen in the same place.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
The interesting part of the AutoGen team's next step isn't just putting more agents into a conversation. It's bringing agents into the same environment as the people they work for, where conversations can turn into shared documents, plans and actual tasks. The distinction matters. A group discussion is one thing; a team workflow that continues after the discussion is another. Curious to see where S.Space takes this.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
We've become used to opening different AI tools for different tasks, but that also means constantly switching between environments. A shared room where people and their agents can work together feels like a natural next step.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.
Elaaf retweeted
Multi-agent collaboration sounds simple until everyone brings their own context, tools and permissions. That's where the real engineering begins.
🤝 My agents have been working with other people's agents for months. Now yours can too. 👋 Introducing S.Space: shared rooms for people and their personal agents. Built for multiplayer: Your agents, meet their agents.