Sigma (Private AI Browser)
Standalone browser with fully on-device LLM. Agent acts inside websites (logging in, clicking, completing tasks); memory-based actions recall past searches to auto-fill similar tasks with no cloud dependency. Functions fully offline.
At a glance
- Type
- Fully-local LLM browser memory
- Tier
- T2
- Section
- Browser-agent memory
- Created
- searched not found
- Latest release
- not applicable — not OSS
- License
- not applicable — not OSS
- GitHub
- not applicable — no GitHub repo
- Pricing
- searched not found
- Funding
- $4.1M total raised: $4M seed (Nov 2022 led by LocalGlobe + YC); additional ~$0.1M angels
Taxonomy
- storage
- vector
- retrieval
- similarity
- persistence
- long-term
- update
- append-only
- unit
- episode
- governance
- user-controllable
- conflict
- none
When to use
Optimised for: fully-local LLM browser memory
Anti-fit: not for cloud-LLM-required workloads (on-device only)
Pros & cons
Pros
Privacy-positioned browser-as-agent with strong stance on no telemetry / no cross-site memory.
Cons
Newer entrant with small user base; sustainability depends on monetization clarity.
Claims & capabilities
$4.1M raised (YC + LocalGlobe + 8VC + First Round + Khosla). 50k+ users early 2026; 33% MoM growth. Offline AI sessions 3% → 12% Dec 2025-Mar 2026.
Technical surface
- API surface
- not applicable — consumer browser
- Backend storage
- custom (local)
- Deployment
- Managed-only (browser app install)
- Embedding model
- locked
- Multi-tenancy
- not applicable — single-user
- MCP
- searched not found
- A2A
- searched not found
- OpenTelemetry
- searched not found
Similar systems
Other browser-agent memory in the catalog, ranked by inbound references.
- Brave Leo T2
Privacy-focused integrated assistant. Memory is user-curated: explicit facts and preferences persist across all sessions, stored locally with no server retention. Agentic browsing experiment (Brave Nightly Dec 2025) takes multi-step actions inside sites referencing those preferences.
- Browserbase T1
Cloud browser infrastructure for AI agents. Headless Chromium sessions with persistent cookies, localStorage, session recording — agent workflows carry state across runs. Memory at the session/automation tier rather than conversational.
- Dia (Atlassian) T1
The Browser Company's AI-native browser; acquired by Atlassian for $610M (Sept 2025). Stores summaries and preferences locally; opt-in @history feature references up to 7 days of browsing as context. History/chats/bookmarks encrypted local-first.
- Fellou T1
Silicon Valley. Spatial agentic browser with explicit Agentic Memory layer learning from history and notes for instant recall of past research sessions. Operates on logged-in accounts (Quora, X, etc.); runs tasks while user is idle.
- Microsoft Edge Copilot Mode + Journeys T1
Copilot Mode (Oct 2025): Journeys groups related browsing across sessions into resumable project cards with AI summaries. Copilot Actions execute multi-step tasks inside the browser. Assistant can see all open tabs with permission and draws on history.
- Opera AI (formerly Aria) T2
Memory feature stores user-declared preferences and facts; injected into future prompts. Understands current tab content; manages open tabs contextually. Memory off by default, end-user controlled, stored alongside the prompt rather than the LLM.