← back to blueprints
knowledge
aminobots kbot
codename · standalone product
Search and reason over all your company knowledge
Knowledge engine with RBAC-scoped search across structured + unstructured data. Multi-agent research, report generation, background indexing. RLHF-tuned with external memory.
580ms p95$0.005 per query88 clear
live · v2.0
architecture
Architecture diagram renders here · click any layer for details. For the live mockup we keep this lightweight — the rendered detail page in production has the full SVG with click-through.
.blueprint.yaml
# Aminobots KBot · live · v2.0 agents: - identifier: distilbert_router - responder: domain_agent_v2 - escalation: human_handoff llm_routing: tier1: deepseek_flash # 80% tier2: qwen_3_6 # 15% tier3: claude_sonnet # 5% safety_gates: limbic_veto: enabled pii_filter: strict audit_trail: all_stages deploy_target: aws_bedrock_agentcore
call the api · python
from aminobots import Blueprint bp = Blueprint.load("kbot") session = bp.run( message="...", user_id="demo_user_001" ) print(session.reply) print(session.trace) # audit + routing
try in browser
[ready] sandbox warming up · response in ~400ms once you submit
< 1s p95
$0.02 of $25.00
pii redacted
deploy to your cloud
aws cdk at mvp · azure + gcp in phase 2
clear scorecard
evaluated against production traffic · anonymised
C
89
L
91
E
87
A
93
R
90
source
· deployed at core-banking client · L1·L2·L3 engineering support
· pairs with NehishKnow on the bfsi side
· license · commercial · aminobots blueprint license