Filter sessions
agents/plan14.82s
sess_01J9X8A4K · 42 spans · $0.164
tools 8retried 1
support/triage3.42s
sess_01J9X8A3P · 7 spans · $0.184
500
chat/assistant6.12s
sess_01J9X8A2E · 14 spans · $0.098
agents/plan22.4s
sess_01J9X8A1Z · 68 spans · $0.312
tools 14
data/extract2.10s
sess_01J9X8A0R · 3 spans · $0.064
copilot/complete0.74s
sess_01J9X89ZY · 1 span · $0.021
Duration
14.82s
LLM calls
6
Tool calls
8
Tokens
54.2k
Cost
$0.164
0s3s6s9s12s15s
plan_run
guardrails · pass
12ms
llm · claude-opus-4.7
1.80s · 8.2k tok
tool · search_tickets
820ms
tool · get_order
640ms · parallel
llm · claude-sonnet-4
1.60s · 12.1k tok
tool · issue_refund
920ms · retried
llm · claude-sonnet-4
2.4s · 18.4k tok
tool · send_email
460ms
llm · claude-sonnet-4 · final
5.4s
output guardrail · pass
Insights · this session
Final LLM call took 5.4s · 36% of the whole run — try streaming the answer back as it's generated.
issue_refund retried once · 480ms added — the tool's p95 is 220ms, consider shorter timeout.
14 tools exposed · only 4 used · prune unused tools from the system prompt to save ~8.4k input tokens per call.
issue_refund retried once · 480ms added — the tool's p95 is 220ms, consider shorter timeout.
14 tools exposed · only 4 used · prune unused tools from the system prompt to save ~8.4k input tokens per call.