Verifiable Trading Decision Agents and Strategy Templates

Guild on 0G 2.0
1

1) Project overview and objectives

What we’re building: TradeOS lets users “vibe-code” trading decision workflows (assets/timeframes/indicators/confirmation + risk rules) in natural language. On 0G, we’ll turn these workflows into verifiable, composable decision agents, where every signal is shipped with provenance (agent spec hash, model/version, data snapshot hash, execution logs) so other apps/protocols can trust and reuse them.

Traction:

  • ~70,000 MAU

  • 23.5% D7 / 18% D30 retention

  • 16,500 pro-created agents

Objectives (Guild MVP):

  • Ship a testnet verifiable signal pipeline (spec → inference → onchain anchoring).

  • Launch a strategy template library (copy/fork/run) backed by verifiable metadata.

  • Provide an API + demo integration for an ecosystem app to consume signals + proofs.

2) Technical architecture and implementation plan

Lean architecture

  • TradeOS App: Strategy builder outputs an Agent Spec (JSON) + schedule/risk constraints.

  • Orchestrator (runtime): Builds context (market data snapshot + features) and dispatches inference tasks.

  • 0G Compute (Inference): Runs LLM/domain inference and returns structured outputs (signal, confidence, rationale, self-checks).

  • 0G Storage: Stores agent specs, templates, logs, backtest summaries; anchors content hashes for auditability.

  • 0G Chain (EVM): Minimal contracts:

    • Agent Registry (spec hash, creator, versioning)

    • Result Anchoring (result hash + references to Storage/DA)

  • 0G DA (optional): Batch-publish hourly/daily signal snapshots for scalable availability.

8-week plan

  • W1–3: Contracts + Storage integration (templates/specs/logs).

  • W4–5: Compute inference integration; structured outputs + provenance.

  • W6–8: Public demo + API; reference integration (or 1 ecosystem pilot); monitoring/rate limits.

3) How we’ll integrate with 0G infrastructure

  • 0G Compute: Execute decision agents and return structured outputs with verifiable metadata.

  • 0G Storage: Persist agent specs/templates/audit logs; keep hashes as canonical verification anchors.

  • 0G Chain: Anchor agent + result hashes onchain so third parties can verify what generated each signal and on which snapshot.

  • 0G DA (if scope fits): Publish batched signal datasets for broad distribution and availability.

4) Team background and experience

  • 10yrs+ global product veteran from Bigtechs and Leading Startups.

  • Solid AI skills, mastered AI pre-training and feature engineering from EB-level data lake

  • Leadership across on-chain AI data infrastructure and on-chain exchange infrastructure—built for real ecosystems, not demos.

  • Deep background in global payment systems and risk/compliance AI, with production reliability + security mindset.

5) Funding requirements and milestones

Requested grant: $30k (plus gas credits / technical review support if available).

Use of funds (lean):

  • 70% Engineering (Compute/Storage integration, contracts, API)

  • 15% Security & QA (contract review, testing, monitoring)

  • 15% Infra & data pipelines

Milestones

  • M1 (Week 3): Agent Registry + Result Anchoring contracts on testnet; Storage-backed template/spec publishing.

  • M2 (Week 5): End-to-end verifiable inference pipeline on 0G Compute; signals + provenance metadata.

  • M3 (Week 8): Public demo + API; reference integration repo (or 1 ecosystem pilot); initial verified template library.