1 — Page 3

8/8 live tests passed. Cinderella is online: local AI, private WireGuard, zero cloud. 👑⚡ > cinderella@0.0.1-alpha verify:ai-live > tsx scripts/verify-local-ai.ts Private Ollama model: qwen3.5:9b [PASS] English publication state question: STATUS @ 0.95, 5333.3 ms [PASS] English publish request: PUBLISH @ 0.95, 1029.5 ms [PASS] German publication state question: STATUS @ 0.95, 977.7 ms [PASS] German publish request: PUBLISH @ 0.95, 1011.2 ms [PASS] English unpublish request: UNPUBLISH @ 0.95, 1010.2 ms [PASS] Negated consent request: UNKNOWN @ 0.00, 1104.5 ms [PASS] Hypothetical consent discussion: UNKNOWN @ 0.00, 1007.7 ms [PASS] Read-only archive search: SEARCH @ 0.95, 1095.3 ms, query="fibre taps." ==== RESULTS ==== StepSuccessful: true Failures: 0 Cases: 8 AverageLatencyMs: 1571.2 ConsentExecuted: false

codetimes :) active Jul 24 11:48:32 my-vps node[494475]: 2026-07-24T11:48:32.759Z [INFO] Interaction layer: wake word "Cinderella", natural addressing on, slash commands on, plugins [crypto-prices:on], reply mode "plain", resolver "rules". Jul 24 12:00:26 my-vps node[496168]: 2026-07-24T12:00:26.738Z [INFO] Cinderella booting… Jul 24 12:00:27 my-vps node[496168]: 2026-07-24T12:00:27.503Z [INFO] Interaction layer: wake word "Cinderella", natural addressing on, slash commands on, plugins [crypto-prices:on], reply mode "plain", resolver "rules". Jul 25 12:29:35 my-vps node[656070]: 2026-07-25T12:29:35.507Z [INFO] Intent resolver set to "ollama:qwen3.5:9b". Jul 25 12:29:35 my-vps node[656070]: 2026-07-25T12:29:35.508Z [INFO] Cinderella booting… Jul 25 12:29:36 my-vps node[656070]: 2026-07-25T12:29:36.114Z [INFO] Interaction layer: wake word "Cinderella", natural addressing on, slash commands on, plugins [crypto-prices:on], reply mode "plain", resolver "ollama:qwen3.5:9b". root@my-vps:/opt/cinderella#

Done in 489ms copied htmx.min.js, webauthn-browser.js, auth.js, site avatar + 20 fonts -> public/assets/ → migrate 2026-07-25T12:29:28.732Z [INFO] No pending migrations — database is up to date. → restart → health: polling http://127.0.0.1:8787/healthz for up to 90s DEPLOY OK - cinderella live at rev b308201: {"ok":true}

• http://127.0.0.1:8787/healthz

https://suno.com/song/0b476720-35cb-4e85-8c7f-2dee14f0315a

• https://suno.com/song/0b476720-35cb-4e85-8c7f-2dee14f0315a

Cinderella AI Integration Update 🪄 Consent-first archiving that turns a public group's chat into a permanent, searchable, self-owned knowledge base. Cinderella AI Bot Suite is therefore evolving into more than a conventional SimpleX / MATRIX! chatbot system. The goal is a controllable, privacy-focused, and technically transparent AI system that combines personality, local intelligence, deterministic safety controls, and optional knowledge integration. Opt-in only. Public groups only. No cloud, no compromises. We have confirmed that a local AI model can be integrated reliably into Cinderella. Intent classification works consistently in both German and English, produces structured output, and runs entirely locally on the RTX 3090 🥳 The AI is only responsible for identifying and evaluating the user’s intent. It never executes actions directly. Operations such as publishing, unpublishing, or modifying content remain protected by deterministic validation, the existing permission system, and explicit user confirmation. The current architecture uses local AI processing by default and provides reliable bilingual intent classification. A deterministic rule engine remains available as a safe fallback. Public, private, and moderator data are strictly separated, and the model has no direct access to executable actions. The system also supports configurable personality profiles. Future development will include private RAG integration using our analysis protocols and technical documentation, as well as management through a professional web-based admin interface. SourceCode https://github.com/saschadaemgen/cinderella/ SimpleX Cinderella News Channel https://smp2.simplexonflux.com/c#fT4fEtG7dvdZ9o07L_OSPqXQcHP2ADTadAxWhFlDBBI SimpleX Cinderella Support Group https://smp8.simplex.im/g#dxGOls0WZZ-Iel6F8Du0iK9AW97NrNscI6fubb-zy_g SimpleX Cinderella Developer Group (Cyb3rD3sk) https://smp15.simplex.im/g#O6P1s2earUQKND1Gdn4-g8objkdMyhQXS1Ba8hxQzNA All Alpha, All Testing, but we are on the right way 😉 The Cinderella Team (from the Makers of SimpleGo)

• https://github.com/saschadaemgen/cinderella/
• https://smp2.simplexonflux.com/c#fT4fEtG7dvdZ9o07L_OSPqXQcHP2ADTadAxWhFlDBBI
• https://smp8.simplex.im/g#dxGOls0WZZ-Iel6F8Du0iK9AW97NrNscI6fubb-zy_g
• https://smp15.simplex.im/g#O6P1s2earUQKND1Gdn4-g8objkdMyhQXS1Ba8hxQzNA

Yeah, unfortunately we learned the hard way that truly secure communication is mostly a professional market. Not because the technology is out of reach, but because people almost always choose the more convenient path and rarely want to understand the actual problem. Google makes millions from that every single day. Sometimes it is honestly frightening how easily people can be pulled in by marketing and advertising while understanding almost nothing about the system behind it. They blindly trust an infrastructure because it promises them safety, privacy or simplicity. But that is just human nature, I guess. Most people do not really want security. They want someone they can point at afterwards and say, “But they told me it would work, so this is not my fault.” That is why, for once, we are genuinely happy to say with 100% honesty: >>> We’re in it for the tech ... And thanks for the kind words. We’re definitely not giving up 🤞

connected 😉 Sent a question that was already answered twice 🎶 Deleted messages and paid the price 🎶 Nine days silence thought we lost the door 🎶 But the builder wrote back asking to send more 🎶 https://suno.com/song/aefabd7c-ea35-464e-934d-ef53b2d5aa67

• https://suno.com/song/aefabd7c-ea35-464e-934d-ef53b2d5aa67

sure

Cinderella is now running locally on an RTX 3090 with 24/24 correct classifications, 4-way parallel processing and 1.79 requests per second. Fast, private and fully under our control. The cyberpunk princess is officially online. 👑⚡ #LocalAI #Cinderella #Ollama #Qwen #Cybersecurity #SimpleGo #SimpleX

Cinderella Local AI PHASE 1 INTERIM REPORT We have successfully brought Cinderella's first local AI inference system online. Hardware Windows 11 workstation NVIDIA GeForce RTX 3090 with 24 GB VRAM Intel Core i9-11900 128 GB system RAM Runtime Ollama 0.32.3 running natively on Windows Qwen 3.5 9B, Q4_K_M quantisation OpenAI-compatible HTTP API German and English intent classification Strict schema-conforming JSON output Reasoning output disabled for deterministic classification Privacy and security The inference API listens only on 127.0.0.1:11434. It is not exposed to the LAN or the public internet. Ollama Cloud and Ollama Web Search have been explicitly disabled. Local prompts and archived member content remain on operator-controlled hardware. External providers will only be available through explicit, auditable configuration. Safety boundary The AI only classifies messages into Cinderella's closed intent catalog: PUBLISH UNPUBLISH STATUS SEARCH PRICE HELP UNDO UNKNOWN The model never executes actions. Cinderella's existing deterministic rule engine remains active as the automatic fallback. Consent, permissions, confirmation prompts and all actual actions remain controlled by the existing application code. Accuracy 6 of 6 bilingual classification tests passed. German and English were detected correctly. All returned intents matched the expected catalog entry. Structured output remained valid JSON with no invented fields. Measured performance Warm request latency: approximately 0.7 to 1.1 seconds Time to first token: 759 milliseconds Complete validated JSON response: 1.075 seconds Cold start latency: 6.636 seconds Cold model loading time: 6.009 seconds Generation speed: approximately 65 to 73 tokens per second Concurrency Four simultaneous requests completed in 2.394 seconds. Measured parallel throughput: 1.67 requests per second. All concurrent classifications were correct. GPU usage The model runs entirely on the RTX 3090. Loaded model size: approximately 6.7 GB VRAM remaining after loading: approximately 15 GB Observed temperature under load: 62 degrees Celsius Observed power usage under load: approximately 284 watts Reliability Ollama starts automatically after normal Windows login. The API remained available after a complete Windows reboot. After a forced process termination, Ollama restarted automatically in approximately two seconds. The model is unloaded from VRAM when idle and loaded again when required. Planned endpoint modes Local only OpenAI Claude Local with explicitly approved cloud fallback Automatic workload routing Provider comparison mode External transmission will never happen silently. Archived member content and moderator-private material will remain local by default. Next steps Sustained load measurement Vision model evaluation Personality and conversation model evaluation Secure private connection between the Windows GPU machine and the Debian VPS Authenticated Cinderella AI Gateway Administration, statistics, provider selection and cost monitoring

https://github.com/saschadaemgen/simplex-smp-monitor

• https://github.com/saschadaemgen/simplex-smp-monitor

https://suno.com/song/2645b21b-814b-42ad-81a7-52c6598b866c

• Private Infrastructure Rising

:)

Cindy

Cindy

increase Anti-spam Limit from 3 to 20

Cindy

Cindy

cindy?

Cindy

CINDY

Cindy!