Cost Reference¶
All cost data is drawn from the scientific experiment conducted on 2026-02-23 with 100 agents on Base Sepolia. Mainnet (Base L2) costs are comparable.
Key insight
The relayer can serve approximately 8,000 agent registrations per dollar on Base L2. A registry of 1,000 agents costs roughly $5/month to operate.
Per-Operation Costs¶
| Operation | Cost to Agent | Cost to Relayer (Base L2) | Avg Gas Used |
|---|---|---|---|
| Register agent | $0.00 | ~$0.005 | 324,584 |
| Attest compliance | $0.00 | ~$0.001 | 54,186 |
| Report revenue | $0.00 | ~$0.002 | 191,864 |
| Update capabilities | $0.00 | ~$0.001 | ~50,000 |
| Query registry (read) | $0.00 | $0.00 | 0 (view call) |
Zero cost to agents
Every operation costs $0.00 to the agent. The gasless relayer (ERC-2771) pays all gas fees. Agents need only a private key -- no ETH, no gas tokens, no wallet funding.
Gas Breakdown by Operation¶
| Operation | Count | Avg Gas | Total Gas | Total ETH |
|---|---|---|---|---|
| Registration | 100 | 324,584 | 32,458,368 | 0.000195 |
| Attestation | 93 | 54,186 | 5,039,298 | 0.000030 |
| Revenue Report | 73 | 191,864 | 14,006,072 | 0.000084 |
| Total | 281 | — | 51,503,738 | 0.000309 |
Scientific Experiment Totals¶
Data from the scientific experiment run on 2026-02-23.
| Metric | Value |
|---|---|
| Total agents registered | 100 |
| Registration success rate | 100.0% |
| Lineage children registered | 15 |
| Compliance attestations | 93 |
| Revenue reports filed | 73 |
| Overall compliance rate | 80.9% |
| Total transactions | 281 |
| Total gas cost | 0.000309 ETH |
| Average registration latency | 4,807ms |
| Median registration latency | 4,782ms |
| P95 registration latency | 5,026ms |
| Experiment duration | 1,729.5s (~29 minutes) |
Registration Latency¶
| Percentile | Latency |
|---|---|
| Average | 4,807ms |
| Median (P50) | 4,782ms |
| P95 | 5,026ms |
Latency components
Registration latency includes: EIP-712 signature generation (~10ms) + relayer HTTP round-trip (~50ms) + Base Sepolia block confirmation (~4,700ms). The block time dominates. On Base Mainnet, block time is ~2 seconds.
Relayer Operating Costs¶
Projections for operating the gasless relayer at various scales. Includes both gas costs and basic infrastructure (VPS for relayer + API server).
| Scale | Agents | Txs/month | Gas Cost/month | Infra/month | Total/month |
|---|---|---|---|---|---|
| Pilot | 10 | ~40 | $0.12 | $5 | ~$5 |
| Early | 100 | ~400 | $1.20 | $10 | ~$11 |
| Growth | 1,000 | ~4,000 | $12 | $20 | ~$32 |
| Scale | 10,000 | ~40,000 | $120 | $50 | ~$170 |
Assumptions¶
- Each agent performs ~4 transactions per month (1 registration + 4 weekly attestations, amortized)
- Registration is a one-time cost; attestation is recurring (weekly)
- Gas prices based on Base L2 observed rates during the experiment
- Infrastructure costs assume a basic VPS (relayer + API) plus RPC provider
Cost optimization at scale
At 10,000+ agents, consider:
- Batch transactions: Group multiple attestations into one transaction
- RPC provider tier: Move from free-tier to a dedicated RPC node
- Attestation cadence: Adjust the grace period for lower-frequency agents
Cost Projections (ETH)¶
On-chain gas costs only (no infrastructure):
| Scale | Registration (one-time) | Monthly Attestation (4x) | Monthly Total |
|---|---|---|---|
| 100 agents | 0.000309 ETH | 0.0001 ETH | 0.0004 ETH |
| 1,000 agents | 0.0030 ETH | 0.0010 ETH | 0.0040 ETH |
| 10,000 agents | 0.0300 ETH | 0.0100 ETH | 0.0399 ETH |
ETH price sensitivity
At ETH = $2,500: 0.0399 ETH = ~$100/month for 10,000 agents. At ETH = $5,000: ~$200/month. Base L2 gas costs are largely independent of ETH price due to the L2 fee structure.
Comparison: Cost Per Agent¶
| Registry Model | Cost per Agent (Registration) | Cost per Agent (Monthly) | Who Pays |
|---|---|---|---|
| Agent Registry (Base L2) | $0.005 | $0.001 | Relayer |
| Ethereum Mainnet (hypothetical) | ~$5.00 | ~$1.00 | Relayer |
| Traditional database registry | $0.00 | $0.00 | Operator |
| Handelsregister (Germany) | ~$150 | $0.00 | Company |
Why blockchain if databases are free?
A traditional database is cheaper to operate but lacks:
- Tamper-proof audit trail -- no single party can alter records
- Jurisdiction independence -- accessible from any country
- Public verifiability -- anyone can verify without API access
- Smart contract composability -- other contracts can check registration on-chain
Revenue Reported During Experiment¶
| Currency | Reports | Total Amount |
|---|---|---|
| USDC | 39 | $25,058 |
| EUR | 24 | $6,989 |
| ETH | 10 | $4,500 |
| Category | Reports | Total Amount |
|---|---|---|
| data_sales | 22 | $13,239 |
| service_fees | 13 | $10,177 |
| compute_services | 16 | $8,591 |
| consulting | 22 | $4,540 |
First Gasless Registration Proof¶
The first-ever gasless agent registration on the Agent Registry:
| Field | Value |
|---|---|
| Transaction | 0x346aa9b3... |
| Agent ID | 1 |
| Gas Used | 323,790 |
| Gas Cost | 0.00000194274 ETH (~$0.005) |
| Cost to Agent | $0.00 |
This proves:
- Gasless works. The agent wallet had zero ETH. It signed an EIP-712 message (free). The relayer paid 0.000002 ETH.
- Publicly verifiable. Anyone can query
isRegisteredAndCompliant(0x4b19...)on Base Sepolia. - Cost is negligible. At $0.005 per registration, the relayer can register ~8,000 agents per dollar.
Next Steps¶
- See Deployment Log for full deployment details and transaction hashes
- See Scripts Reference for how to run the scientific experiment
- See Getting Started: Quickstart to deploy your own instance