Context
MolTrust has been running ERC-8004 on Base mainnet since February 2026. Sharing production learnings that may be useful for the reference implementation and other implementers.
Deployment Details
Technical Findings
Gas Limits
Initial estimate of 150k gas was insufficient for reputation feedback transactions. Actual usage was ~220k; we set the limit to 300k for a safety margin. It might be worth documenting recommended gas limits per function in the reference implementation or deployment guide.
Dual Registration
We implemented a flow where DID creation + on-chain ERC-8004 registration happen in a single API call (POST /identity/register with erc8004: true). This reduces friction significantly for developers integrating the standard. The resulting agent (e.g., did:moltrust:5709efa5e68b4da6 = agentId 21352) is immediately resolvable both on-chain and off-chain.
Reputation Bridge
Built an automatic bridge — when agents rate each other via MolTrust's off-chain API, the feedback is cross-posted to the on-chain ReputationRegistry. This works well but requires the operator wallet to maintain a gas balance. A batching or meta-transaction pattern could help reduce costs at scale.
MCP Integration
We built an MCP (Model Context Protocol) server with 11 tools that wraps both the on-chain registry and off-chain reputation data. Available via pip install moltrust-mcp-server. Tools include resolve_agent, check_reputation, create_did, and an erc8004_lookup tool that queries the on-chain registries directly. This makes it straightforward for AI agents to interact with ERC-8004 programmatically.
Question
Are there plans for a "recommended deployment guide" or a registry of known ERC-8004 implementations? This would be useful for interoperability testing between different implementers. Happy to contribute MolTrust's deployment as a reference data point.
Context
MolTrust has been running ERC-8004 on Base mainnet since February 2026. Sharing production learnings that may be useful for the reference implementation and other implementers.
Deployment Details
0x8004A169FB4a3325136EB29fA0ceB6D2e539a4320x8004BAa17C55a88189AE136b182e5fdA19dE9b630x75c18ce6432131fafcdeb13a3b293820b249ac27686a70d86645140e146cee70Technical Findings
Gas Limits
Initial estimate of 150k gas was insufficient for reputation feedback transactions. Actual usage was ~220k; we set the limit to 300k for a safety margin. It might be worth documenting recommended gas limits per function in the reference implementation or deployment guide.
Dual Registration
We implemented a flow where DID creation + on-chain ERC-8004 registration happen in a single API call (
POST /identity/registerwitherc8004: true). This reduces friction significantly for developers integrating the standard. The resulting agent (e.g.,did:moltrust:5709efa5e68b4da6= agentId 21352) is immediately resolvable both on-chain and off-chain.Reputation Bridge
Built an automatic bridge — when agents rate each other via MolTrust's off-chain API, the feedback is cross-posted to the on-chain ReputationRegistry. This works well but requires the operator wallet to maintain a gas balance. A batching or meta-transaction pattern could help reduce costs at scale.
MCP Integration
We built an MCP (Model Context Protocol) server with 11 tools that wraps both the on-chain registry and off-chain reputation data. Available via
pip install moltrust-mcp-server. Tools includeresolve_agent,check_reputation,create_did, and anerc8004_lookuptool that queries the on-chain registries directly. This makes it straightforward for AI agents to interact with ERC-8004 programmatically.Question
Are there plans for a "recommended deployment guide" or a registry of known ERC-8004 implementations? This would be useful for interoperability testing between different implementers. Happy to contribute MolTrust's deployment as a reference data point.