SignalPR
AI-native PR execution platform for founders: real-time media event radar, newsjacking angle generation, journalist vector RAG matching, and outreach sync to Instantly, Smartlead, and Apollo.
Identity + memory sidecar for agents you already run — validated personality, episodic recall, and a hybrid API so your LLM keeps generation.
ensure → prepare → your LLM → complete
SoulOS owns persona, memory, and MSV drift. Your app keeps Bedrock, OpenAI, LiteLLM, or any chat stack.
Idempotent bootstrap with external_key and a .soul.json.
POST /hybrid/prepare returns a ready system_prompt + memories.
Stream tokens with your existing client — SoulOS stays on embeddings when you want.
Ingest the turn and optionally reflect MSV with POST /hybrid/complete.
Pick the path that matches your stack.
Independent products use SoulOS for persistent persona and episodic memory. Listing is factual, not endorsement.
AI-native PR execution platform for founders: real-time media event radar, newsjacking angle generation, journalist vector RAG matching, and outreach sync to Instantly, Smartlead, and Apollo.
Interactive digital legacy platform: AI-guided life-story interviews, encrypted voice archives, family natural-language Q&A, and a digital twin that answers from the narrator's own recordings.
Family Croatia travel expert: kid-friendly itineraries, hidden beaches, islands, national parks, and AI trip planning curated by Vedrana.
AI press release distribution and journalist outreach for founders: match live reporter requests (Qwoted, HARO, and more), draft newsroom-ready releases, and distribute to press sites without an agency retainer.
Common questions for humans and answer engines — same content as the FAQPage structured data.
SoulOS is an open-source identity and episodic memory sidecar for AI agents. It provides HEXACO MSV personality, pgvector memory, and a hybrid prepare/complete API so your existing LLM keeps generation.
Use SoulOS when you need persistent persona beyond a static system prompt, episodic memory across sessions, a hybrid sidecar next to Bedrock/OpenAI/LiteLLM, MCP tools for memory and identity in Cursor or Claude, or ready MIT SoulPack personas (travel, sales, tutor, support, and more).
ensure_avatar → POST /hybrid/prepare → your LLM → POST /hybrid/complete. See the sidecar integration guide and npm run smoke:hybrid.
Yes. The kernel, SDK, Studio, SoulPacks, and examples are MIT-licensed. The project site is free on GitHub Pages and synced from the same repository.
SoulPacks are 23 first-party MIT persona packages in packs/soulpacks/ (travel, SDR, tutor, tech support, developer coach, exec assistant, research, customer success, security, PM, data, recruiter, content, and more). List with GET /v1/soulpacks, import with POST /v1/avatars/import-soulpack, or use the Studio SoulPacks tab.
On the GitHub Pages catalog at /soulpacks/ — search by name, role, or tag, then open a detail page for SOUL, identity, style, and a copy-paste import example. Each pack is also listed in the site sitemap for crawlers.
Open-source MIT kernel you can inspect, fork, and self-host.