Founder at @simpliciomcp, AI Labs: simpleti.com.br/ Microsaas: ligne.site
Guarulhos, SP, Brazil
- Tweets 2,267
- Following 2,025
- Followers 558
- Likes 13,922
Herói no meio da multidão
Um herói de capa roxa e luvas amarelas circula pela multidão à noite. Vídeo criado com inteligência artificial.
#IA #Shorts #RealOficial
2
121
59
Herói da escadaria
Um herói de capa roxa desfila pela escadaria colorida. Vídeo criado com inteligência artificial.
#IA #Shorts #RealOficial
1
148
Tudo começou com um DUVIDO
Tudo começou com um "DUVIDO". Vídeo criado com inteligência artificial.
#IA #Shorts #RealOficial
84
Me ajuda @acgfbr @paulogrego_
Não estou conseguindo agendar posts e nem colocar na fila do Launch os vídeos criados por AI
Peguei um vídeo, troquei o rosto, queria publicar em todas as minhas redes sociais de uma vez
Ele só reconhece cortes, nao videos Xablau
#xablau
92
Lesson: parallelize review, serialize merges. Squads caught bugs a single reviewer missed: bwrap killed by a seccomp filter, tests that would silently flip, a module that wouldn't import in prod. Numbers measured, not estimated. 5/5
38
The numbers: 374 s -> 93 s per merge. 228 s -> 146 s of review per PR. 315k -> 75k tokens of coordinator context. So far: 19 PRs merged, 10 issues closed, 8 Haiku->Sonnet escalations, 4 bugs blocked pre-merge. 4/5
1
43
The setup: Watcher, Turbo & planner, Tools & dashboard. Each squad has a lead that reviews, runs tests vs main and mutation-checks the fix. One owner per file and a fixed merge order keep conflicts out. 3/5
1
50
Before: one coordinator reviewed and merged every PR alone, while 29 workers queued behind it. After: 3 squad leads review in parallel and a general coordinator only merges what they approve. Same agents, no bottleneck. 2/5
1
168
We reorganized the AI agents working on simplicio-loop: from 1 coordinator + workers to squads. Measured from agent run metadata: -75% time per merge, -36% review time per PR, -76% coordinator context, 3x reviewers in parallel. 1/5
Haiku 5.5 is now available in the Claude Platform and Claude Code. On average, it costs around 75% less to run than Haiku 4.5.
It pairs well with Opus 5.5 or Sonnet 5.5 as a subagent. Use it for high-volume, cost-sensitive tasks like summaries, compactions, or database queries.
1
2
725
$$ 50k - Fiz essa Automação e coloquei dinheiro no bolso:
Contexto: TRansportadora tinha 3 funcionarias gastando 4 horas por dia para enviar canhotos das entregas no sistema financeiro
Agora com Automacao demora 20 segundos cada canhoto..
Veja a ultima imagem.. é a melhor
1
1
139
Olha ai eu construindo o JEV Local antes do JEV, o @simpliciomcp
#bolhadev #bolhasec @BolhaDevs @BolhaSec @sseraphini
1
65
Pra votar tenho que preencher isso tudo? Porra, até desisti, man, coloca um login com google e já era.. que burocracia @heliotsx
1
2
409
Resultados vão chegar
Acesse: simpleti.com.br/simplicio
#bolhadev #bolhasec @BolhaDevs @BolhaSec @sseraphini
2
5
240
💥 $10 por $1
Pessoal, terminei meu motor de execução do @simpliciomcp , ajustado..
fluxo de velocidade e economia
pqp, 5 dias nisso
olha o resultado:
#bolhadev #bolhasec @BolhaDevs @sseraphini @acgfbr @heliotsx
5
207
@acgfbr @heliotsx Big dog, coloca um selecionar todos aqui nesse pop-up, na tela de agendamento de posts ... Please ...
Economize ate 96% de tokens com @simpliciomcp
1
90
@acgfbr @heliotsx Ontem fiz uma apresentacao de 2 horas sobre AI, agora to usando Real Oficial para gerar os cortes, serão 75 cortes .. wow
Use minha ferramenta que economiza até 96% de tokens! @simpliciomcp
1
1
4
2,818
To terminando Software que economiza ate 96% tokens qulquer llm por sessao @simpliciomcp , vou terminar o app desktop e falo com vcs , meu usuario beta test passou de $38.5 para $1.35 por dia com minha solução
1
1
32
I woke up; I had left the GPT 5.6 Sol Ultra Fast running all night.
00:58 - 73%
06:47 - 65%
An 8% drop in usage over 5 hours of continuous operation.
@sseraphini @acgfbr @manodeyvin @lucas_montano @oraulsena @thisabri @Teknium @theo @marclou #buildinpublic #software #claude #codex #grok #gemini #bolhadev #bolhasec #openrouter #hermes #qween #kimi #deepseek #token #github #elon #ai #dev
1
1
2
294
Os tokens não expiram durante uma hora de uso contínuo com o GPT 5.6 Sol xHigh.
@sseraphini @acgfbr @manodeyvin @lucas_montano @oraulsena @thisabri @Teknium @theo @marclou #buildinpublic #software #claude #codex #grok #gemini #bolhadev #bolhasec #openrouter #hermes #qween #kimi #deepseek #token #github #elon #ai #dev
🇧🇷 Brasil faz parte disso!
Muito feliz em fazer parte dos contribuidores ativos do Hermes Agent
Foram mais de 40 PRs aprovados no Hermes Agent
5 PRs do Bot Mode - 1 modulo inteiro isolamento de Grupos dos Agents
Vamos pra cima!
#bolhadev #bolhasec @sseraphini #hermes @Teknium @NousResearch
1
1
14
1,999
Tokens do not expire during one hour of continuous use with GPT 5.6 Sol xHigh.
@sseraphini @acgfbr @manodeyvin @lucas_montano @oraulsena @thisabri @Teknium @theo @marclou #buildinpublic #software #claude #codex #grok #gemini #bolhadev #bolhasec #openrouter #hermes #qween #kimi #deepseek #token #github #elon #ai #dev
🇧🇷 Brazil is part of this!
Very happy to be among the active contributors to Hermes Agent.
Over 40 PRs approved for Hermes Agent.
5 PRs for Bot Mode—including an entire module for Agent Group isolation.
Let's do this!
#hermes @Teknium @NousResearch
3
7
1,046
I opened Group Agents for Hermes Bot Mode: saved teams, isolated histories, explicit routing, lead-first replies, private sessions, bounded context, and async-storage safety.
github.com/NousResearch/Herm…
#bolhadev #bolhasec #hermes @Teknium
1
1
155
I opened a durable messaging-protocol fix for Hermes Bot Mode: custom SOUL files preserve the handoff protocol and the Agent Inbox bootstrap stays tied to the live roster.
github.com/NousResearch/Herm…
#bolhadev #bolhasec #hermes @Teknium
96
I opened an Active Profile pane for Hermes Bot Mode, bringing the agent’s identity, provider, model, description, skills, toolsets, and edit action into one focused view.
github.com/NousResearch/Herm…
#bolhadev #bolhasec #hermes @Teknium
2
162
I opened a Hermes Bot Mode Cronjob fix: when the active profile is already the target agent, routines execute directly instead of recursively delegating to themselves.
github.com/NousResearch/Herm…
#bolhadev #bolhasec #hermes @Teknium
1
122
I opened a resilience fix for Hermes Bot Mode: a temporary profiles.list failure no longer clears the Bot roster—the last healthy state stays visible while recovery is clear.
github.com/NousResearch/Herm…
#bolhadev #bolhasec #hermes @Teknium
84
I'm going to launch a product that saves over 98% of tokens with any LLM, check out the benchmark:
1
52
🔥 A maioria dos agentes de IA entrega apenas um *patch*... e para por aí.
Seu *backlog* continua crescendo.
Você ainda precisa supervisionar cada *Pull Request* (PR).
Tokens são desperdiçados à toa.
Autonomia real: zero.
A solução?
🔁 simplicio-loop
Resolve todo o seu *backlog* enquanto você dorme.
descobrir → implementar → verificar → fazer *merge* → fechar → monitorar 24/7
Funciona com QUALQUER LLM.
Protegido por barreiras de segurança robustas + mecanismo de interrupção de custos (*kill-switch*).
Até 96% menos consumo de tokens.
Não é um *chatbot*.
É um agente executor.
👉 github.com/wesleysimplicio/s…
#IA #Agentes #IAAutônoma #LLM #DevTools #OpenSource #AgenticAI
1
1,465
🔥 Most AI agents hand you a patch… and stop.
Your backlog keeps growing.
You still babysit every PR.
Tokens burn for nothing.
Zero real autonomy.
The solution?
🔁 simplicio-loop
Finishes your entire backlog while you sleep.
discover → implement → verify → merge → close → watch 24/7
Works on ANY LLM.
Behind hard safety gates + cost kill-switch.
Up to 96% fewer tokens.
Not a chatbot.
A worker.
👉 github.com/wesleysimplicio/s…
#AI #Agents #AutonomousAI #LLM #DevTools #OpenSource #AgenticAI
1
1
3
166
🔥 Você dá a qualquer LLM uma tarefa de codificação de uma linha...
Ele alucina arquivos.
Pula os testes.
Quebra o código existente.
Precisão?
Frequentemente abaixo de 40%.
A solução?
📷 simplicio-dev-cli
Transforme uma tarefa de uma linha em uma mudança de código verificada: → Contexto de mapeamento → Contrato de 6 camadas → Diff limpo → Testes reais → Recibo de evidência
Funciona com Claude, DeepSeek, Codex, Gemini, Hermes, Cursor... qualquer LLM. Até 99% de precisão.
Sem necessidade de fine-tuning.
📷github.com/wesleysimplicio/s…
#AI #CodingAgents #LLM #DevTools #OpenSource #AgenticAI
5
1,571
🔥 You give any LLM a one-line coding task...
It hallucinates files.
Skips the tests.
Breaks existing code.
Accuracy? Often under 40%.
The solution?
⚡ simplicio-dev-cli
Turn a one-line task into a verified code change:
→ Mapper context
→ 6-layer contract
→ Clean diff
→ Real tests
→ Evidence receipt
Works with Claude, DeepSeek, Codex, Gemini, Hermes, Cursor... any LLM.
Up to 99% accuracy. No fine-tuning needed.
👉 github.com/wesleysimplicio/s…
#AI #CodingAgents #LLM #DevTools #OpenSource #AgenticAI
3
1
5
856
🇺🇸 🔥 Every AI agent arrives WITHOUT a map.
Every task starts from ZERO.
Tokens burned. Context lost. Pure chaos.
The fix?
🗺️ simplicio-mapper
Turns ANY repo into a deterministic map.
Context becomes an ARTIFACT — not another bloated prompt.
Scan once → reuse forever.
Agents that actually KNOW your code.
👉 github.com/wesleysimplicio/s…
#AI #Agents #LLM #OpenSource #DevTools
2
2
210
Simplicio Agent
- 10x faster than Hermes Agent
- 90% more cost-effective with Simplicio Runtime
1
1
74
🚨 Most AI agents stop when they think they’re done.
Simplicio Loop stops only when the work is actually done. ✅
🔁 Discover work
🧠 Understand requirements
🗺️ Build dependency graphs (DAG)
⚡ Autoscale agent fleets
💻 Implement code
🧪 Run tests
🌐 Verify real behavior
🔥 Fix failures automatically
👀 Resolve review comments
🚀 Open PRs
✅ Merge changes
📌 Close issues with evidence
♾️ Repeat 24/7
What makes it different?
🧩 6 specialized skills
🌎 11 runtimes
🔌 43 extension points
💰 Up to 96% fewer tokens
🛡️ Safety gates
🔍 Evidence-driven completion
🤖 Autonomous software delivery
Works across:
⚡ Claude Code
⚡ Codex
⚡ Cursor
⚡ GitHub Copilot
⚡ Gemini
⚡ OpenCode
⚡ Kiro
⚡ Aider
⚡ Hermes
⚡ OpenClaw
⚡ And more…
Most AI tools generate code.
Simplicio Loop delivers software.
From:
“Finish all open issues.”
To:
Merged PRs, passing CI, resolved reviews, closed tickets, verified results.
No fake “done”.
No blind automation.
No endless prompting.
Just autonomous execution. 🚀
#AI #Agents #ClaudeCode #Codex #Cursor #OpenSource #LLM #Automation #SoftwareEngineering #DevTools #AgenticAI #GitHub #Programming #AIEngineering #SimplicioLoop #HermesAgent 🚀🔥♾️
2
1
121
🚨 Most AI agents stop when they think they’re done.
Simplicio Loop stops only when the work is actually done. ✅
🔁 Discover work
🧠 Understand requirements
🗺️ Build dependency graphs (DAG)
⚡ Autoscale agent fleets
💻 Implement code
🧪 Run tests
🌐 Verify real behavior
🔥 Fix failures automatically
👀 Resolve review comments
🚀 Open PRs
✅ Merge changes
📌 Close issues with evidence
♾️ Repeat 24/7
What makes it different?
🧩 6 specialized skills
🌎 11 runtimes
🔌 43 extension points
💰 Up to 96% fewer tokens
🛡️ Safety gates
🔍 Evidence-driven completion
🤖 Autonomous software delivery
Works across:
⚡ Claude Code
⚡ Codex
⚡ Cursor
⚡ GitHub Copilot
⚡ Gemini
⚡ OpenCode
⚡ Kiro
⚡ Aider
⚡ Hermes
⚡ OpenClaw
⚡ And more…
Most AI tools generate code.
Simplicio Loop delivers software.
From:
“Finish all open issues.”
To:
Merged PRs, passing CI, resolved reviews, closed tickets, verified results.
No fake “done”.
No blind automation.
No endless prompting.
Just autonomous execution. 🚀
#AI #Agents #ClaudeCode #Codex #Cursor #OpenSource #LLM #Automation #SoftwareEngineering #DevTools #AgenticAI #GitHub #Programming #AIEngineering #SimplicioLoop #HermesAgent 🚀🔥♾️
3
1
3
240
🚨 Most AI agents stop when they think they’re done.
Simplicio Loop stops only when the work is actually done. ✅
🔁 Discover work
🧠 Understand requirements
🗺️ Build dependency graphs (DAG)
⚡ Autoscale agent fleets
💻 Implement code
🧪 Run tests
🌐 Verify real behavior
🔥 Fix failures automatically
👀 Resolve review comments
🚀 Open PRs
✅ Merge changes
📌 Close issues with evidence
♾️ Repeat 24/7
What makes it different?
🧩 6 specialized skills
🌎 11 runtimes
🔌 43 extension points
💰 Up to 96% fewer tokens
🛡️ Safety gates
🔍 Evidence-driven completion
🤖 Autonomous software delivery
Works across:
⚡ Claude Code
⚡ Codex
⚡ Cursor
⚡ GitHub Copilot
⚡ Gemini
⚡ OpenCode
⚡ Kiro
⚡ Aider
⚡ Hermes
⚡ OpenClaw
⚡ And more…
Most AI tools generate code.
Simplicio Loop delivers software.
From:
“Finish all open issues.”
To:
Merged PRs, passing CI, resolved reviews, closed tickets, verified results.
No fake “done”.
No blind automation.
No endless prompting.
Just autonomous execution. 🚀
#AI #Agents #ClaudeCode #Codex #Cursor #OpenSource #LLM #Automation #SoftwareEngineering #DevTools #AgenticAI #GitHub #Programming #AIEngineering #SimplicioLoop #HermesAgent 🚀🔥♾️
3
1
922
Introducing Hermes Agent + Simplicio — FREE 🔥
Why spend 100% of your tokens when your agent can run smarter?
Simplicio brings smart orchestration, neural cache, and compressed context to AI agent workflows — saving up to 96% of tokens.
⚡ Faster execution
💸 Lower costs
🧠 Smarter context
🚀 Leaner workflows
Run smarter. Spend less. Build more.
github.com/wesleysimplicio/s…
#AI #LLM #CodingAgents #HermesAgent #Simplicio #GitHub
92
Introducing Le Gros Chaton + Simplicio — FREE 🔥
The big kitten with a tiny token bill.
Why burn millions of tokens when your AI workflow can run smarter?
Simplicio brings smart orchestration, neural cache, and compressed context to Le Gros Chaton workflows — saving up to 96% of tokens.
🟢 Up to 96% token savings
⚡ Faster responses
💸 Lower costs
🧠 Smarter context
Le Gros Chaton + Simplicio = leaner AI execution.
#AI #LLM #MistralAI #LeGrosChaton #Simplicio #CodingAgents
50
隆重推出 Simplicio — 免费 🔥
既然你的智能体可以更聪明地运行，
为什么还要白白消耗数百万个 tokens？
🟢 Simplicio：节省 96%
🟡 RTK：节省 80%
🟠 Caveman：节省 65%
智能编排。神经缓存。压缩上下文。
#AI #LLM #CodingAgents #GitHub
1
93
Logo logo disponível para todos
Claude Fable 5 mesmo resultado, menos tokens
#bolhadev #bolhasec @sseraphini @dionisiodev @lucas_montano
1
1
4
870
