File size: 3,662 Bytes
dc7cc4d 0b8db51 dc7cc4d 0b8db51 dc7cc4d 0b8db51 dc7cc4d 0b8db51 dc7cc4d 0b8db51 dc7cc4d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 | import gradio as gr
import requests
import os
# Read backend URL from Hugging Face secret
BACKEND_URL = os.environ.get("KERNL_BACKEND_URL", "").rstrip('/')
if not BACKEND_URL:
BACKEND_URL = None
def query_kernl(scenario, with_brain):
"""Call the Kernl /agent/handle endpoint and format the response."""
if not BACKEND_URL:
return "❌ Backend URL not configured. Please set the KERNL_BACKEND_URL secret."
if not scenario or not scenario.strip():
return "❗ Please enter a scenario."
try:
response = requests.post(
f"{BACKEND_URL}/agent/handle",
json={
"company_id": "rivanly-inc",
"scenario": scenario,
"with_brain": with_brain
},
timeout=30
)
response.raise_for_status()
data = response.json()
output = []
output.append(f"**Action:** `{data.get('action', 'N/A')}`")
output.append(f"**Rule Applied:** `{data.get('rule_applied', 'N/A')}`")
output.append(f"**Message:** {data.get('message_to_customer', data.get('answer', 'N/A'))}")
if data.get('evidence'):
output.append(f"**Evidence:** {data.get('evidence')}")
output.append(f"**Skill Matched:** `{data.get('skill_matched', 'N/A')}`")
output.append(f"**Confidence:** `{data.get('confidence', 'N/A')}`")
return "\n\n".join(output)
except requests.exceptions.ConnectionError:
return "❌ Cannot connect to Kernl backend."
except requests.exceptions.Timeout:
return "❌ Request timed out."
except Exception as e:
return f"❌ Error: {str(e)}"
# Theme
theme = gr.themes.Soft(
primary_hue="teal",
secondary_hue="teal",
neutral_hue="gray",
font=gr.themes.GoogleFont("Inter")
)
with gr.Blocks(theme=theme, title="Kernl – Operational Memory for AI Agents") as demo:
gr.Markdown("""
# 🧠 Kernl
### Operational memory for AI agents
Kernl compiles how your company actually decides things – from Slack, SOPs, and tickets – into an executable skills file.
Any agent. Any task. Correct every time.
""")
with gr.Row():
with gr.Column(scale=1):
scenario_input = gr.Textbox(
label="Enter your business scenario",
placeholder="Example: Enterprise customer, 18 months tenure, wants $1,200 refund",
lines=4
)
with_brain_toggle = gr.Checkbox(
label="🧠 Use Company Brain (Kernl)",
value=True,
info="ON = Kernl uses compiled company knowledge. OFF = generic AI answer."
)
submit_btn = gr.Button("Ask Kernl", variant="primary", size="lg")
with gr.Column(scale=2):
output_box = gr.Markdown(label="Kernl's Response", value="*Your answer will appear here...*")
gr.Markdown("""
---
### Try these example scenarios (copy & paste):
- `Enterprise customer, 18 months tenure, wants $1,200 refund`
- `Annual plan customer, day 10 of subscription, $300 refund requested`
- `Customer reporting P0 bug on dashboard, enterprise plan`
- `Customer showing 3 churn signals in last 30 days`
- `Startup requesting 40% discount`
---
**Built with AMD MI300X, vLLM, and LangGraph** | [GitHub](https://github.com/your-repo) | **Track 1: AI Agents & Agentic Workflows**
""")
submit_btn.click(
fn=query_kernl,
inputs=[scenario_input, with_brain_toggle],
outputs=output_box
)
demo.launch() |