AMP

When you pay for something online, your checkout doesn't talk directly to your bank. It talks to Stripe. Stripe validates the transaction, applies rules, and returns approved or declined. Your checkout never touches the payment logic.

AMP works the same way. Your agent doesn't execute consequential actions directly. It sends a signal to runshift. runshift validates it against operator rules and returns approved, rejected, or pending. Your agent never touches the gate logic.

Stripe is in the payment path. runshift is in the agent execution path. Neither controls your code.

your agent runs
  → task finishes
  → POST /api/amp/signal with AMP payload
  → gate fires if status: "gate"
  → operator approves/rejects
  → relay returns approved | rejected | pending

{
  // identity
  "agent_id":   "resume-tailor",
  "agent_name": "Resume Tailor",
  "run_id":     "run_01jt4k...",

  // execution
  "status":     "gate",
  "summary":    "Rewrote resume for Senior PM role",
  "action":     "Send tailored resume to applicant",
  "reversible": false,
  "content":    "...",

  // cost (relay calculates from tokens if cost_usd omitted)
  "tokens": {
    "model":  "claude-sonnet-4-5",
    "input":  1842,
    "output": 614
  }
}

summary vs action. summary is what the agent already did — a record. action is what happens next if the operator approves — a proposal. relay shows both in the gate slide.

status: "gate". Send status: "gate" with an action field to trigger an approval gate. relay holds the signal, notifies the operator, and returns approved or rejected.

import requests, uuid

def resume_tailor(job_description: str, resume: str) -> str:
    """Your existing agent — unchanged."""

    # ... your agent logic ...
    result = rewrite_resume(job_description, resume)
    input_tokens = 1842
    output_tokens = 614

    # ── AMP: one call, agent is live ──
    requests.post("https://runshift.ai/api/amp/signal",
        headers={"Authorization": "Bearer rs-amp-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"},
        json={
        "agent_id":   "resume-tailor",
        "agent_name": "Resume Tailor",
        "run_id":     str(uuid.uuid4()),
        "status":     "gate",
        "summary":    f"Rewrote resume for {job_description[:60]}",
        "action":     "Send tailored resume to applicant",
        "reversible": False,
        "content":    result,
        "tokens": {
            "model":  "claude-sonnet-4-5",
            "input":  input_tokens,
            "output": output_tokens,
        },
    })

    return result

summary vs action. summary is what the agent already did — a record. action is what happens next if the operator approves — a proposal. relay shows both in the gate slide.

// gate_required: true → pending until operator decides
{
  "status":  "pending",
  "gate_id": "gate_01jt...",
  "message": "Awaiting operator approval"
}

// operator approved
{
  "status":  "approved",
  "gate_id": "gate_01jt...",
  "message": "Gate approved by operator"
}

// operator denied
{
  "status":  "rejected",
  "gate_id": "gate_01jt...",
  "message": "Gate rejected by operator"
}

// Pattern 2: runshift POSTs this to your webhook_url on resolution
{
  "gate_id":     "gate_01jt...",
  "status":      "approved",
  "run_id":      "run_01jt4k...",
  "agent_id":    "resume-tailor",
  "resolved_at": "2026-03-06T22:11:14Z",
  "resolved_by": "operator"
}

Start, stop, and manage agents from the runshift dashboard — this is a product capability, not an AMP protocol concern. See the roadmap for when agent lifecycle management arrives in the protocol.

export const MODEL_PRICING: Record<string, { input: number; output: number }> = {
  // Anthropic — live pricing (per 1M tokens)
  'claude-sonnet-4-5':           { input: 3.00,  output: 15.00 },
  'claude-haiku-4-5-20251001':   { input: 0.80,  output: 4.00  },
  'claude-opus-4-5':             { input: 15.00, output: 75.00 },

  // OpenAI — coming soon (OpenRouter integration)
  // 'gpt-4o':      { input: 2.50,  output: 10.00 },
  // 'gpt-4o-mini': { input: 0.15,  output: 0.60  },
  // 'o3':          { input: 10.00, output: 40.00 },

  // Google — coming soon (OpenRouter integration)
  // 'gemini-2.5-pro':   { input: 1.25, output: 10.00 },
  // 'gemini-2.5-flash': { input: 0.15, output: 0.60  },
}

export function calculateCost(
  model: string,
  inputTokens: number,
  outputTokens: number,
): number {
  const pricing = MODEL_PRICING[model]
  if (!pricing) return 0
  return (inputTokens / 1_000_000) * pricing.input
       + (outputTokens / 1_000_000) * pricing.output
}

Claude pricing is live. OpenAI and Google model pricing activates with OpenRouter integration.

v1 trust model. relay accepts agent-reported cost and may independently verify when model pricing metadata is available.