selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Agora Cosmica logo

Self-host Agora Cosmica

v1.3.1

Living library of 30 historical figures you can talk to

ReleaseOpen SourceAGPL-3.0
Updatedlast week
LanguageTypescript
Setup difficultyModerate
MobileResponsive
Stars
149
RAM
32 GB unified RAM on Apple Silicon, or 16 GB VRAM on NVIDIA for local LLM
Ports
8080 (main web application), 8880 (Kokoro English TTS), 8000 (Whisper speech-to-text), 8887 (Qwen3-TTS German TTS (nvidia profile))
Docker
ghcr.io/chipmates/agoracosmica
Install with Docker Compose
Story chapter paused with ask bar showing typed question and Echo's answer as textFacts panel for a story chapter showing date, place, documented vs recreated content, and sources

What is Agora Cosmica?

Agora Cosmica is a nonprofit, open-source education platform presenting 30 historical figures spanning 2,500 years. Each figure offers six learning formats: narrated Story chapters (~13 minutes each), Wisdom teachings for live conversation, Prism dialogues with three other thinkers, Quest Socratic challenges, Free Talk open conversation, and Council four-voice debates. All content is bilingual (English/German) with pre-recorded audio totaling nearly 80 hours in English and over 85 in German. The self-hosted Docker Compose stack includes the web app, Kokoro English TTS, Whisper transcription, and optional Qwen3-TTS for German. Operators can configure the app to use any OpenAI-compatible LLM (LM Studio, Ollama, vLLM, llama.cpp) so conversations never leave their hardware. User data stays in the browser's encrypted IndexedDB; no accounts, no tracking cookies. The code is AGPL-3.0; content moves to CC-BY 4.0 within 6-12 months after the May 2026 release.

Should you self-host Agora Cosmica?

Agora Cosmica scores 48/100 on selfhost.directory (how scoring works) and ranks #14 of 30 in Education. It has 149 GitHub stars.

It shipped 7 stable releases in the past year. The latest, v1.3.1, came out on 8 September 2026. New versions land roughly every 2 weeks.

Setup is moderate: expect some configuration beyond a single command. An official Docker image is available. It is AGPL-3.0-licensed and written mostly in Typescript.

Its closest alternative on selfhost.directory is CTFd (70/100, 6 893 stars). It scores higher overall, so compare the two before you pick. See Agora Cosmica vs CTFd side by side.

Features

  • 30 Historical Figures
  • Bilingual English/German
  • Content Safety & Crisis Resources
  • Fact-Check Panels
  • Live AI Conversation (Echo)
  • Local Mode (Offline LLM)
  • Local Speech & Transcription
  • Pre-recorded Audio Library
  • Privacy-first architecture
  • Six Learning Formats
  • WCAG 2.2 AA Accessibility

Agora Cosmica README

How to install Agora Cosmica

Install Agora Cosmica with Docker Compose

Environment variables
AGORA_HOST_PORT, AGORA_MEDIA_BASE_URL, AGORA_TTS_KOKORO_PORT, AGORA_STT_PORT, AGORA_TTS_QWEN_PORT, AGORA_AUDIO_API_URL

git clone https://github.com/chipmates/agoracosmica.git cd agoracosmica docker compose up -d

Open http://localhost:8080. On first run, paste an OpenRouter API key when prompted.

Default stack includes app (port 8080), Kokoro English TTS (port 8880), Whisper STT (port 8000). Use `--profile nvidia` for German Qwen3-TTS. Apple Silicon users can run native MLX TTS/STT via provided scripts.

How to configure Agora Cosmica

  • AGORA_AUDIO_API_URLOptional centralized audio backend (OpenAI-compatible /v1/audio/speech and /v1/audio/transcriptions)

    Optional centralized audio backend (OpenAI-compatible /v1/audio/speech and /v1/audio/transcriptions)

  • AGORA_HOST_PORT8080Host-side port for the app

    Host-side port for the app

  • AGORA_MEDIA_BASE_URLhttps://media.agoracosmica.orgContent CDN for figure prompts, voice profiles, and pre-recorded audio

    Content CDN for figure prompts, voice profiles, and pre-recorded audio

  • AGORA_STT_PORT8000Host-side port for Whisper (transcription)

    Host-side port for Whisper (transcription)

  • AGORA_TTS_KOKORO_PORT8880Host-side port for Kokoro (English speech)

    Host-side port for Kokoro (English speech)

  • AGORA_TTS_QWEN_PORT8887Host-side port for Qwen3-TTS (German, nvidia profile)

    Host-side port for Qwen3-TTS (German, nvidia profile)

Agora Cosmica releases & version history

FAQ

Is it available in German?

Yes. Everything is fully bilingual — English and German — across the content, the interface, and the audio.

Is it open source?

Yes. The code is open under AGPL-3.0, so anyone can read it, audit the privacy claims, contribute, or run their own copy. Content moves to CC-BY within 6-12 months.

How do you handle my privacy?

Conversations are stored only in your browser, never on servers. No tracking cookies, no third-party analytics, no profiles. If you bring your own key, it never touches the project's servers.

Can I run it entirely on my own machine?

Yes. Local Mode lets you point the app at any OpenAI-compatible LLM (LM Studio, Ollama, vLLM) and run the audio stack in Docker. With Local Mode and self-host, no conversation, voice, or text data leaves your machine.

Do I need an account to start?

No account needed. Open the app and start learning immediately with 30 free daily messages. No email or signup required.

Does it cost anything?

You get 30 free messages a day with no signup. For more, bring your own OpenRouter key. The platform is nonprofit and open source with no ads or paywalls.

Is this really the historical person speaking?

No. Every figure is an AI Echo — a voice shaped by their own writing and ideas. It draws on their real philosophy but stays an interpretation, never a recording or the actual person.

What is Agora Cosmica?

An education platform built as a living library where 30 historical figures teach through narrated life stories, core teachings, four-voice debates, and live AI conversation in English and German.

More in Local AI

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…