Row-Level Security policies in PostgreSQL enforce tenant isolation. Every connection sets app.tenant_id. Application code cannot bypass tenant boundaries.
Explicit task checklists replace rigid loops. Tool requests require justification (what/how/why). Sequential and parallel execution modes. Real-time progress via GraphQL subscriptions.
Vector search runs directly in PostgreSQL using pgvector with HNSW indexes. Configurable embedding models generate high-dimensional vectors for text and images. Per-tenant tables with Row-Level Security. MinIO object storage for multimodal retrieval.
Bare-metal architecture with full infrastructure ownership. FastAPI + GraphQL API, Celery workers for async execution, PostgreSQL with pgvector and Row-Level Security, Redis for pub/sub, MinIO for object storage, and HashiCorp Vault for secrets. Declarative infrastructure with NixOS + Ansible.
Deploy on your own bare-metal servers or VMs. NixOS modules provision PostgreSQL+pgvector, Redis, MinIO, HashiCorp Vault, and HAProxy. Ansible playbooks for fleet management. Complete control over data residency and infrastructure with zero cloud vendor lock-in.
Run locally with Docker Compose or `etherion up`. PostgreSQL, Redis, and MinIO containers with named volumes. Python 3.11+ virtual environment. One-command bootstrap with `etherion bootstrap`.
Systemd service management. Alembic database migrations. HashiCorp Vault for secrets injection. HAProxy + Nginx for load balancing and DDoS protection. Patroni for PostgreSQL HA. FRRouting for BGP failover.
From goal to execution in three simple steps
Start by describing your business objective in plain English. Go beyond simple tasks—assign complex, multi-step goals and provide the Orchestrator with the strategic context it needs.
Your private Orchestrator analyzes the goal, creates a plan, and assembles a team of specialist agents. Watch the entire reasoning process unfold in our transparent, Grok-style execution trace.
The Orchestrator synthesizes the work of its specialist team into a single, cohesive final output that directly achieves your goal. Provide feedback to make your workforce even smarter over time.
Based on my analysis of your top three competitors...
Bare-metal architecture with database-enforced multi-tenancy, asynchronous job execution, and real-time updates. Full infrastructure ownership with NixOS + Ansible.
OAuth-secured connectors ingest data from your tools into PostgreSQL with pgvector. Configurable embedding models enable semantic search. All data is tenant-isolated with Row-Level Security.
IO performs dual search (KB + web), evaluates teams, and enforces fail-closed tool approval. TeamOrchestrator executes the 2N+1 loop: N specialist agents work in parallel, each validating tool requests with what/how/why justification against the ToolManager registry. A final synthesis step integrates all findings into a coherent response.
Jobs run in the background using Celery workers and Redis as the message broker. Two worker pools handle different workloads: worker-agents for orchestration, worker-artifacts for ingestion and heavy processing. Real-time status updates stream via GraphQL subscriptions.
All tools use Model Context Protocol (MCP) and connect to third-party systems via OAuth. OAuth tokens are encrypted in HashiCorp Vault. Tool calls validate against the ToolManager registry, require pre-approval for the team, and for write operations, require explicit user confirmation. Rate limiting via token bucket + Redis prevents API abuse.
Every artifact agents create is stored in MinIO and indexed in PostgreSQL. Documents, datasets, code, and media are searchable and retrievable. Full execution traces are archived as JSONL for replay and audit.
Row-Level Security policies in PostgreSQL enforce tenant isolation at the database layer. Every connection sets app.tenant_id. Application bugs cannot cause cross-tenant data leaks.
PyMuPDF extracts text and images from PDFs. Configurable embedding models generate high-dimensional vectors for both text and images. All embeddings are stored in PostgreSQL with pgvector HNSW indexes for fast cosine-distance search. Files stored in MinIO with per-tenant buckets.
All tool requests require what/how/why justification. Requests are validated in 4 steps: (1) Is it registered in ToolManager? (2) Pre-approved for this team? (3) Are tenant credentials available? (4) For write operations, confirmed by user? Blueprint creation validates tools against the registry—no hallucinated tools can enter production. Fail-closed policy ensures every tool invocation is auditable and secure.
The Team Orchestrator selects execution mode based on task complexity. Sequential mode runs one specialist at a time. Parallel mode (future) will run all specialists concurrently. Mode selection is logged in execution trace events.
One specialist active at a time. Tool requests handled immediately. Checklists maintained throughout execution. Current default mode.
All specialists run concurrently. Tool requests queued and processed in FIFO order. Deferred for future release.
Every job execution is recorded with complete LangChain message lists, tool IO, and specialist delegations. Traces are archived to MinIO as JSONL and indexed in PostgreSQL for semantic search. Replay artifacts enable 100% reconstruction of any past execution.
JWT-based authentication with invite-only onboarding. OAuth tokens encrypted in HashiCorp Vault. Subdomain validation enforces 8 rules, reserves 90+ system subdomains, and blocks 1662 banned words. Users cannot switch tenants after signup.
Run locally with `etherion up` or Docker Compose. PostgreSQL, Redis, and MinIO containers with named volumes. Python 3.11+ virtual environment. One-command bootstrap and teardown with Etherion CLI.
Agent definitions are stored in PostgreSQL with tenant isolation. Each agent has a system prompt, tool allowlist, model configuration, and execution limits. The Platform Orchestrator creates agent team blueprints through conversational interaction in the Agents Foundry UI, which is strictly separated from task execution in the Threads Dashboard.
MCP tools connect agents to external systems through OAuth. All write operations require explicit confirmation. Rate limiting and circuit breakers prevent API abuse. Tool definitions are database-driven with tenant-scoped allowlists.
Etherion is open source under the MIT License. The complete codebase, infrastructure modules, and documentation are available on GitHub. Deploy on your own bare-metal servers with full control over your data and infrastructure.
Full source code access. Fork, modify, and deploy without restrictions.
Deploy on your own servers. NixOS modules + Ansible playbooks included for complete infrastructure as code.
Agent-first contribution model. Complete technical documentation and development logs included.
Core platform shipped with fully local execution. Now moving toward general availability.
The foundational infrastructure, the 2N+1 Orchestrator, and the core agentic framework are built and validated.
All services wired end-to-end (orchestrator, async workers, MCP toolchain, knowledge base, repository). Production-like environments validated, security hardened, observability instrumented.
Etherion ships as fully self-contained Python packages on PyPI (etherion and etherion-tui). Install and run the entire platform locally — no cloud account required. The terminal UI provides one-command setup, OAuth provider management, agent orchestration, and live process monitoring, all backed by your own PostgreSQL, MinIO, and Redis stack.
Full public launch with managed cloud hosting, team collaboration, and enterprise SLA options — alongside continued support for fully self-hosted deployments.
Etherion was born from a single, powerful insight: the future of AI is not just about building better tools, but about fundamentally changing our relationship with work. As a self-taught, curious, and fast-thinking systems architect, My vision is to empower users to move beyond tedious implementation and focus on what humans do best: high-level strategy and creativity. Etherion is the culmination of this vision A Platform designed not just for coders, but for architects.