Local proxy enforces your custom policies + DLP before traffic leaves the machine.
Loading routing…
Switch to Direct mode
Direct mode points OpenClaw straight at the AceTeam Gateway, bypassing this local proxy. You lose:
Custom policies (the rules in the dashboard's Custom Policies section)
DLP / secret-leak scanning
Local audit log + cost dashboard
You gain ~one network hop of latency.
Edit ~/safeclaw/config/openclaw.json and replace http://aep-proxy:8899/v1 with https://aceteam.ai/api/gateway/v1 on both openai and anthropic providers, then restart OpenClaw:
To return to Safe mode, swap the URLs back and restart again.
Total Cost
$0.000000
Calls
0
Safety Status
Budget
Policies
Policy Tester
Evaluate a natural-language policy against arbitrary text, without
running an agent. Pick one of the rules defined above, paste in a
sample, and see whether it passes or gets flagged.
Pick a policy and enter text, then press Test.
Safety Signals
Governance Context
Attestation (Signed Verdicts)
Call Timeline
Policy Compliance
100%
Threats Blocked
0
Total Agent Spend
$0.00
Total Calls
0
Cost Savings
Calls blocked before reaching LLM0
Estimated cost saved$0.00
Average cost per call$0.00
Compliance Status
Safety Breakdown
Calls passed0
Calls flagged0
Calls blocked0
Cost by Entity
Data Governance
Attestation
Safety verdicts recorded for all
calls
Every PASS/FLAG/BLOCK decision is logged with timestamp and
detector IDACTIVE
Cryptographic signing
(Ed25519)
Signed verdict chain for tamper-proof audit
trailROADMAP
Third-party attestation
Independent verification of safety enforcement
claimsROADMAP