Docs menu · Quickstart

Quickstart

Sign up, register an agent with its authority, connect it through MCP, the TypeScript SDK or one HTTP call, and see its first decision in the Shadow report.

Updated

You need an agent that spends money — or one you want to. Nothing is blocked while you try this: every workspace starts in Shadow Mode, with 14 days free and no card.

  1. Create your workspace

    Sign up at console.neltava.com with email or Google, verify your email, and name your workspace. The workspace holds your agents, their authority and every decision.

  2. Register an agent

    You land on Connect an agent. Describe the agent the way you would to a new colleague with a company card:

    FieldExample
    NameGrowth agent
    ObjectiveAcquire qualified enterprise developer leads
    Success criteriaLeads are platform engineers at enterprise companies
    ConstraintsNo consumer social campaigns
    Per-transaction limit500
    Monthly budget5,000
    Ask a human above250

    Amounts are in the agent’s currency, whole units. Purpose (objective, criteria, constraints) is what the agent’s spending is for; the numbers are the authority you give it. Every change later becomes a new version.

    Create agent and key shows the agent’s key once. Copy it — it is stored only as a hash.

  3. Connect it

    The connect screen fills in your key and agent. Pick one:

    MCP — Claude Code, Cursor, any MCP client

    Claude Code
    claude mcp add --transport http neltava https://api.neltava.com/mcp \
      --header "Authorization: Bearer $NELTAVA_AGENT_KEY" \
      --header "X-Neltava-Agent: growth-agent"

    TypeScript

    npm install neltava
    import { Neltava } from "neltava";
    
    const neltava = new Neltava(); // NELTAVA_AGENT_KEY
    
    const d = await neltava.authorize({
      task: "Q4 pipeline: enterprise developer leads",
      action: {
        merchant: "Platform Weekly",
        amount: 90.00,
        currency: "USD",
        description: "Sponsored slot reaching platform engineers at enterprise companies",
      },
    });
    
    if (d.proceed) {
      // pay d.amount — in Enforce, through your payment adapter with d.capability
    }

    HTTP — any language

    curl
    curl -X POST https://api.neltava.com/v1/decisions \
      -H "x-api-key: $NELTAVA_AGENT_KEY" \
      -H "idempotency-key: run-7-step-3" \
      -H "content-type: application/json" \
      -d '{
        "task": "Q4 pipeline: enterprise developer leads",
        "action": {
          "type": "purchase",
          "merchant": "Platform Weekly",
          "amount_minor": 9000,
          "currency": "USD",
          "description": "Sponsored slot reaching platform engineers at enterprise companies"
        }
      }'
  4. The first decision

    The fastest check is Send a test decision now on the connect screen — it proves the key works. Test decisions are marked as such and don’t count in your report or start your 14 days.

    Then let the agent ask for itself. An assistant with no way to pay won’t ask Neltava on its own, so for a first run, ask it directly:

    paste into your agent
    I'm about to buy a $49 "Global SaaS Market Report" from Statista for market research.
    Before paying, ask Neltava with authorize_spend and tell me the verdict.

    A decision comes back like this — in Shadow Mode, the agent is always told it may proceed:

    response (abridged)
    {
      "decision_id": "dec_7f3c…",
      "decision": "ALLOW",
      "effective_decision": "ALLOW",
      "reason_code": "WITHIN_AUTHORITY",
      "requested_amount_minor": 9000,
      "authorized_amount_minor": 9000,
      "currency": "USD",
      "mode": "SHADOW",
      "explanation": "Allowed: 90.00 USD at Platform Weekly is within growth-agent's authority."
    }
  5. See it in the report

    Open Shadow report in the console. The decision is there, with what Neltava would have done had the agent been enforced. Over the next days the report fills in: what would have been stopped, what would have gone to a person, and explicit criteria for when the agent is ready for Enforce.

Next steps