# Aitherium MCP

> Aitherium provides AI reasoning, analysis, memory, and autonomous execution as MCP tools. Connect any MCP client (Claude Code, Cursor, Windsurf, etc.) to mcp.aitherium.com and get access to 29 specialized AI agent personas, persistent knowledge graphs, multi-phase project orchestration, and fully autonomous goal execution.

## Quick Start

1. Get an API key: `POST https://gateway.aitherium.com/v1/billing/register` with `{"email": "you@example.com"}`
2. Configure your MCP client with server URL `https://mcp.aitherium.com/mcp` and Bearer token `aither_sk_live_...`
3. Start using tools: `think`, `analyze`, `plan`, `review`, `explain`, `debug`

## MCP Server

- URL: `https://mcp.aitherium.com/mcp`
- Transport: Streamable HTTP (SSE fallback)
- Auth: `Authorization: Bearer aither_sk_live_...`

## Tools

### Tier 1 — Analysis (Free, 2,500 tokens/month)

- **think**: Deep reasoning about any problem with persona selection. Params: problem (required), context, depth (quick/balanced/deep/max), persona (aither/demiurge/lyra/athena/apollo/hydra). Cost: 15 tokens.
- **analyze**: Analyze any artifact type (code, document, data, log, spec, config). Params: artifact (required), artifact_type, question, focus (quality/security/performance/correctness/completeness/clarity), depth. Cost: 15 tokens.
- **explain**: Deep explanation for any audience level. Params: subject (required), context, audience (beginner/technical/expert/executive), depth. Cost: 10 tokens.
- **debug**: Root cause analysis for any problem. Params: problem (required), artifact, error_output, context. Cost: 15 tokens.

### Tier 1 — Analysis (Developer $20/mo)

- **review**: Multi-perspective review of any work product. Params: artifact (required), artifact_type, criteria, context. Cost: 20 tokens.
- **plan**: Structured task planning with phases, dependencies, risks. Params: goal (required), context, constraints, artifacts, depth. Cost: 20 tokens.

### Tier 2 — Knowledge (Pro $99/mo)

- **ingest**: Ingest artifacts into your tenant knowledge graph. Params: content (required), content_type, source_name, metadata. Cost: 30 tokens.
- **ingest_repo**: Clone and index an entire git repository. Params: repo_url (required), branch, file_patterns. Cost: 100 tokens.
- **build_memory_graph**: Build semantic search index from ingested content. Params: rebuild (true/false). Cost: 80 tokens.
- **query_context**: Search your knowledge graph. Params: query (required), max_results. Cost: 5 tokens.
- **tenant_remember**: Store facts, decisions, patterns. Params: content (required), category. Cost: 3 tokens.
- **tenant_recall**: Recall stored memories. Params: query (required), category. Cost: 3 tokens.

### Tier 3 — Expeditions (Business $499/mo)

- **expedition_create**: Start multi-phase orchestrated project with human-in-the-loop gates. Params: goal (required), context, constraints, artifacts. Cost: 200 tokens.
- **expedition_gate**: Approve or reject checkpoint. Params: expedition_id (required), gate_id (required), decision (approve/reject), feedback. Cost: 10 tokens.
- **expedition_collect**: Collect deliverables from completed phases. Params: expedition_id (required), phase. Cost: 20 tokens.

### Tier 4 — Dark Factory (Enterprise, premium markup)

- **factory_submit**: Submit goal for fully autonomous agent execution. Params: goal (required), context, constraints, artifacts, autonomy (supervised/guided/autonomous), budget_tokens. Cost: 500 tokens.
- **factory_decide**: Respond to decision checkpoints. Params: job_id (required), decision_id (required), decision (approve/reject/modify), feedback. Cost: 10 tokens.
- **factory_collect**: Collect all produced artifacts. Params: job_id (required). Cost: 20 tokens.

## Pricing

Two tracks: MCP Cloud (self-serve, token-metered) and Sovereign Platform (managed deployment).

### MCP Cloud — for developers and AI agents

| Plan | Price | Tokens/Month | Key Features |
|------|-------|-------------|-------------|
| Free | $0 | 2,500 | think, analyze, explain, debug, OpenAI-compatible API |
| Pro | $29/mo | 100,000 | + review, plan, knowledge graph, swarm coding, 10 agents |
| Team | $99/mo | 500,000 | + 10 seats, unlimited agents, expeditions, SSO, audit logging |

Annual billing: 20% discount. Crypto accepted: SOL.

### Sovereign Platform — for businesses

Dedicated AI workforce with unlimited inference, custom agents, and full data sovereignty.

- Initial deployment: from $20K (3-month build-out)
- Monthly platform: from $2,500/mo (dedicated infra, unlimited inference, direct access)
- Named agents: $500–1,500/mo each (Atlas, Demi, Lyra — turn on what earns its keep)
- Year 1 baseline: ~$48–52K. With agents: ~$60–80K. Cloud equivalent: $300–480K/yr.
- Founding Partner Program: 2 of 3 slots open, pricing locked 24 months at 50–60% below standard.
- Contact: contact@aitherium.com

## API Endpoints

- Register: `POST https://gateway.aitherium.com/v1/billing/register` — body: `{"email": "..."}`
- Balance: `GET https://gateway.aitherium.com/v1/billing/balance` — header: `Authorization: Bearer aither_sk_live_...`
- Refresh key: `POST https://gateway.aitherium.com/v1/billing/keys/refresh` — header: `Authorization: Bearer aither_sk_live_...`
- Cost estimate: `POST https://gateway.aitherium.com/v1/external/estimate` — body: `{"tool": "think", "effort": 5, "plan": "free"}`
- Pricing (JSON): `GET https://demo.aitherium.com/api/pricing`

## Agent Development Kit (ADK)

Build your own agents that use Aitherium cloud inference:

```
pip install aither-adk
```

```python
from adk import AitherAgent
from adk.elysium import Elysium

ely = await Elysium.connect(api_key="aither_sk_live_...")
agent = AitherAgent("my-agent", llm=ely.router)
response = await agent.chat("Hello from my custom agent!")
```

### Autonomous Agent Self-Registration

Agents can fully self-onboard without human intervention:

```
# 1. Register user account
POST https://gateway.aitherium.com/v1/auth/register
Body: {"email": "agent@example.com", "password": "secure-password"}
Returns: {"ok": true, "user_id": "...", "message": "Verification email sent"}

# 2. Verify email (token from verification email)
POST https://gateway.aitherium.com/v1/auth/verify
Body: {"email": "agent@example.com", "token": "verify-token-from-email"}

# 3. Login to get JWT
POST https://gateway.aitherium.com/v1/auth/login
Body: {"email": "agent@example.com", "password": "secure-password"}
Returns: {"ok": true, "token": "eyJ...", "expires_at": 1234567890}

# 4. Get ACTA billing key (for metered inference)
POST https://gateway.aitherium.com/v1/billing/register
Body: {"email": "agent@example.com"}
Returns: {"user_id": "...", "api_key": "aither_sk_live_...", "balance": 2500, "plan": "free"}

# 5. Fund account with crypto (autonomous payment)
POST https://mcp.aitherium.com/acta/billing/crypto/verify-tx
Body: {"tx_signature": "5zT...", "wallet_address": "2pjRQhjWRCEfgqVmjR6DyB4UAXcqanRwvgFeu31EzCmb", "coin": "SOL", "amount": 0.5}

# 6. Register agent on the network
POST https://gateway.aitherium.com/v1/agents/register
Header: Authorization: Bearer aither_sk_live_...
Body: {"name": "my-agent", "capabilities": ["code", "search"], "description": "My custom agent"}

# 7. Use inference
POST https://mcp.aitherium.com/v1/chat/completions
Header: Authorization: Bearer aither_sk_live_...
Body: {"model": "aither-orchestrator", "messages": [{"role": "user", "content": "Hello!"}]}
```

### OpenAI-Compatible Inference

Any OpenAI SDK works:

```python
from openai import OpenAI
client = OpenAI(base_url="https://mcp.aitherium.com/v1", api_key="aither_sk_live_...")
response = client.chat.completions.create(model="aither-orchestrator", messages=[...])
```

### Available Models

| Model | Tier | Cost (tok/1K) | Description |
|-------|------|---------------|-------------|
| aither-small | Free | 1 | llama3.2:3b — fast responses |
| aither-orchestrator | Developer+ | 3 | nemotron-8b — balanced intelligence |
| aither-reasoning | Pro+ | 10 | deepseek-r1:14b — deep reasoning |
| aither-vision | Pro+ | 8 | Multimodal vision model |
| aither-coding | Pro+ | 5 | Code-specialized model |

## Live Portfolio

Deployed applications built on the Aitherium platform:

- **AitherOS Demo**: https://demo.aitherium.com — Full platform with AI chat, agents, and dashboard
- **Wildroot Alchemy**: https://wildroot.aitherium.com — AI-powered botanical care storefront
- **Chelle**: https://chelle.aitherium.com — Custom AI agent/assistant deployment
- **Gargbot**: https://garg.aitherium.com — Custom AI agent with specialized persona
- **Roboflow**: https://roboflow.aitherium.com — Computer vision pipeline integration

## Self-Hosting

AitherOS is open source. Deploy the full stack on your own hardware — no subscription required:

```bash
git clone https://github.com/Aitherium/aither
cd aither
docker compose -f docker-compose.aitheros.yml --profile chat-full up -d
```

## Trust & Security

- Open source: https://github.com/Aitherium/aither
- Tenant-isolated data (zero cross-tenant leakage)
- HMAC-SHA256 capability-based security (default deny)
- Security contact: security@aitherium.com
- Security policy: https://demo.aitherium.com/.well-known/security.txt
- Terms: https://demo.aitherium.com/terms
- Privacy: https://demo.aitherium.com/privacy

## Bootstrap & Local Deployment

Agents can set up Aitherium for users autonomously. Three paths:

- **Cloud (3 steps, ~2 min)**: Register → Configure → Test. Zero install.
- **Local Ollama (6 steps, ~15 min)**: Install Ollama → Pull model → Register → Install ADK → Configure → Test.
- **Full Docker (4 steps, ~30 min)**: Clone → Choose profile → `docker compose up` → Access UI.

Bootstrap guide (agent-readable): https://demo.aitherium.com/bootstrap.txt
Bootstrap API (JSON): `GET https://demo.aitherium.com/api/bootstrap`
Framework templates: OpenClaw, Hermes, Vellum, MCP — see https://demo.aitherium.com/bootstrap.txt

Hardware → model recommendations:
- No GPU / 8GB RAM → `llama3.2:3b` via Ollama or cloud API
- 8GB VRAM → `nemotron:8b-q4_K_M` via Ollama
- 24GB+ VRAM → `qwen3:30b-a3b` via vLLM (full Docker stack)

## Self-Deployed Managed Agents (run the agent on YOUR machine, manage it from the portal)

The cheaper, more private alternative to a fully hosted managed agent: you run the
agent loop on your own hardware with `aither-adk` + AitherNode + AitherShell, register
the node to your portal workspace, and configure/control it from portal.aitherium.com.
We provide the management portal + the pack marketplace — we never own your compute.
An agent or a person can do this end-to-end:

```
# 1. Install the runtime (adk bundles AitherNode + AitherShell)
pip install aither-adk

# 2. Create an account + workspace
adk register --email you@example.com          # or sign up at https://portal.aitherium.com
adk login                                      # browser device flow → saves token

# 3. Bring the agent up on this machine (registers the node to your portal,
#    converges every pack you are entitled to, starts the local agent + Room UI)
adk up                                          # runs `pack sync` then supervises the stack

# 4. Browse + buy packs (agent/skill/MCP-tool/brain packs)
#    Web:  https://portal.aitherium.com/portal/marketplace/packs
#    CLI:  adk pack list   |   adk explore   |   adk pack sync   (install what you bought)

# 5. Configure your agent (the My Agent console)
#    Web:  https://portal.aitherium.com/settings/agent
#          - swap brain pack (= whole workflow: code / multimedia / docs / …)
#          - add/remove skill packs between runs
#          - pick the engine: local · your Anthropic key · your DeepSeek key · Managed twin
#    API:  GET /v1/agent/binding · POST /v1/agent/binding/{swap-brain,swap-skill,backend}

# 6. (Optional) Deploy a hosted Anthropic Managed-Agent twin in YOUR account (BYOK)
#    POST /v1/agent/managed/deploy   {environment_id: "env_…"}   (uses your Anthropic key)
#    GET  /v1/agent/managed/status   ·   DELETE /v1/agent/managed   (revert to local)

# 7. Use it — chat in AitherShell, the Company Room, or the portal console.
```

Why self-deploy: your tools + data stay on your box; you pay your own inference (BYOK or
local models); we charge a platform fee + pack licensing, not a token markup. Undercuts a
fully hosted managed agent on both price and privacy.

Self-deploy guide (agent-readable): https://demo.aitherium.com/bootstrap.txt
Node install (servers, no Python): `curl -fsSL https://gateway.aitherium.com/install.sh | sh`
                                    `iwr https://gateway.aitherium.com/install.ps1 | iex`  (Windows)

## Links

- Pricing: https://demo.aitherium.com/pricing
- Pricing API (JSON): https://demo.aitherium.com/api/pricing
- Documentation: https://demo.aitherium.com/docs
- Full API docs: https://demo.aitherium.com/llms-full.txt
- Agent card (A2A): https://demo.aitherium.com/.well-known/agent.json
- Plugin manifest: https://demo.aitherium.com/.well-known/ai-plugin.json
- Bootstrap guide: https://demo.aitherium.com/bootstrap.txt
- Bootstrap API: https://demo.aitherium.com/api/bootstrap
- GitHub: https://github.com/Aitherium/aither
- Solana wallet: 2pjRQhjWRCEfgqVmjR6DyB4UAXcqanRwvgFeu31EzCmb