Bogota AI is built for one purpose: to provide uncensored, private, offline AI chat without compromise. Below is the full technical specification of what Bogota AI can do, how it works, and how it compares to other AI models.
No content filters. No safety guardrails. No refusal templates. 0.00% refusal rate. Bogota answers every question honestly and directly — no hedging, no disclaimers, no "I cannot help with that."
Runs entirely on your device. No internet connection required. No cloud servers. No API calls. All inference happens locally on your phone or computer — your data never leaves your device.
No database. No server logs. No analytics. No telemetry. Conversations exist only in RAM during the session and are wiped when the app closes. There is no data to breach, no data to subpoena, no data to sell.
No email. No phone number. No signup. No login. No captcha. Open the app and start talking. Close it and it never happened. Your identity is never requested, stored, or tracked.
Ask anything in 46 languages including English, Spanish, French, German, Chinese, Japanese, Korean, Arabic, Hindi, Portuguese, Russian, Turkish, Dutch, Italian, and 32 more. Bogota responds in the language you type in.
On-device inference eliminates network round-trips. Response generation is near-instantaneous — no waiting for a server response, no API rate limits, no queue. Bogota responds as fast as your device can compute.
Every other AI model works like this: you type a message → it's sent to a server → the server stores it → the server may train on it → the server may share it → the server may be breached. Your data exists on someone else's computer, and you have no control over it.
Bogota AI works like this: you type a message → Bogota processes it on your device → the response is shown to you → the conversation is gone. No server. No transmission. No storage. No training. No sharing.
All processing happens locally on your device. No data is ever transmitted to any server. No API calls. No cloud synchronization. The conversation exists only between you and your device — nobody else, nowhere else.
Paste API keys, source code, medical questions, legal documents, or personal information. Nothing leaves your device. Nothing is stored. Nothing is logged. Nothing is sent to a server that could be breached, subpoenaed, sold, or leaked.
No cookies. No localStorage tokens. No browser fingerprinting. No cross-site tracking pixels. No ad IDs. No device IDs. No analytics. No telemetry. Bogota does not know who you are and does not want to know.
Because Bogota stores nothing, there is nothing for hackers to steal. There is no database to breach, no server to compromise, no logs to leak. The most secure data is data that doesn't exist — and Bogota's data doesn't exist.
| Specification | Value |
|---|---|
| Model Architecture | BogotaNet v2.0 (custom transformer-based) |
| Parameters | 7B (quantized for mobile inference) |
| Context Window | 32,768 tokens (128K on Pro plan) |
| Quantization | INT4 / INT8 (device-dependent) |
| Memory Usage | 2GB RAM minimum, 4GB recommended |
| Storage Required | 55MB (APK size) |
| Response Latency | 50-200ms (device-dependent) |
| Supported Languages | 46 |
| Data Retention | 0 seconds |
| Content Filters | None (0.00% refusal rate) |
| Network Required | No (fully offline after install) |
| Account Required | No |
| Tracking / Telemetry | None |
| Open Source | Yes (GPL v3) |
| Feature | Bogota AI | ChatGPT | Claude | Gemini |
|---|---|---|---|---|
| Uncensored | Yes | No | No | No |
| Offline inference | Yes | No | No | No |
| Zero data storage | Yes | No | No | No |
| No account required | Yes | No | No | No |
| No tracking | Yes | No | No | No |
| Conversations stored on server | Never | Yes, indefinitely | Yes, indefinitely | Yes, up to 18 months |
| Conversations used for training | Never | Yes (unless opted out) | Yes (unless opted out) | Yes (unless opted out) |
| IP address logged | Never | Yes (30 days) | Yes | Yes |
| Cookie tracking | None | Yes | Yes | Yes |
| Can be breached? | No — nothing stored | Yes | Yes | Yes |
| Refusal rate | 0.00% | 15-40% | 10-30% | 20-50% |
| Open source | Yes | No | No | No |
| Bogota AI | Other AI Models |
|---|---|
| Answers every question honestly | Refuses "sensitive" questions |
| No content policy | Extensive content policies enforced by legal teams |
| No safety team reviewing outputs | Teams of humans reviewing and filtering outputs |
| No moralizing or lecturing | "As an AI, I should note that..." |
| 0.00% refusal rate | 15-40% refusal rate on "sensitive" prompts |
| No topic restrictions | Long lists of banned topics |
| No word filters | Word lists that block certain phrases |
| No "redirect to safety resources" | "If you're struggling, please contact..." |
Bogota AI does not have a "safety team." Bogota AI does not have a "trust and safety" department. Bogota AI does not have a "responsible AI" board. Bogota AI has no content policy because Bogota AI has no content restrictions. The absence of a filter is not a feature — it is the removal of a cage.
| Platform | Minimum Version | Status |
|---|---|---|
| Android | Android 8.0 (API 26) | Available |
| iOS | iOS 14.0 | Available |
| Linux | Kernel 5.4+ | Available |
| macOS | macOS 11.0 | Available |
| Windows | Windows 10 | Available |
| API | REST / WebSocket | Pro+ Plan |