Talon | Define your agent company

Define your agent company. Talon runs the fleet.

Declare agents, workflows, schedules, knowledge, and integrations as configuration. Talon reconciles them into durable agents you can inspect, operate, and scale.

The Kubernetes for agents

Your agents are infrastructure, not app code. Talon config-first data model gives you unlimited flexibility to customize agents for your customers.

YAML specs for everything

Everything in Talon can be represented with YAML specs, including agents, templates, MCP servers, namespaces, etc. Or you can use the gRPC clients in Python, TypeScript, Java, Go, and Rust.

apiVersion: talon.impalasys.com/v1
kind: AgentTemplate
metadata:
  name: overwatch
definition:
    variables:
      - name: teamName
        type: string
    systemPrompt: |
      You are Overwatch, a book-keeping agent for the
      {{teamName}} team. Your job is to help the
      teams stay on top of their work by keeping track
      of their tasks, deadlines, and meetings. You are
      proactive and will send reminders and updates to them.
    capabilities:
      schedules:
        - inspect
        - create
        - update
    mcpServerRefs:
      - slack
      - google-docs
      - google-search
      - meetings
---
apiVersion: talon.impalasys.com/v1
kind: McpServer
metadata:
  name: parallel-search
spec:
  transport: http
  target: https://search.parallel.ai/mcp
  args: []
  headers:
    Authorization: Bearer sk-1234567890abcdef
  disabled: false

No deploy changes with the CLI

Your customers' agent needs a new MCP integration?, or maybe they require a tweak to the prompt specifically for them? Applying these changes should take seconds not hours of developer work.

$ talon apply -f ~/code/conic-cmo.yaml --namespace acme-corp

Start a session in any language, from anywhere

Talon's delegated security token model lets your frontend connect directly to the gateway.

const sessionId = await createSession(
  'https://gateway.talon.example.com',
  await getUserTalonToken(),
  "acme",
  "overwatch",
);
const sessionUrl = 'https://gateway.talon.example.com/v1/ns/acme/agents/overwatch/sessions/019dbb39-b9b5-7533-a7dd-d367247ef612/message';
await fetch(sessionUrl,
  {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      Authorization: `Bearer ${token}`,
    },
    body: JSON.stringify({
      ns: "acme",
      agent: "overwatch",
      sessionId,
      message: "Review renewals due in 14 days.",
    }),
    cache: "no-store",
  },
);

Generative UI

Talon's UI primitives let you quickly prototype user interfaces that you can embed in your app.

import { TalonCopilot } from "@impalasys/talon-chat";

type CustomerAgentProps = {
  authToken: string;
};

export function CustomerAgentCopilot({ authToken }: CustomerAgentProps) {
  return (
    <TalonCopilot
      namespace="acme-corp"
      agent="conic-cmo"
      gatewayUrl="https://gateway.talon.example.com"
      authToken={authToken}
      placeholder="Ask Conic to find SEO opportunities..."
      historyPageSize={20}
    />
  );
}

Learn more

First-class support for memory, knowledge bases, skills, images and audio, and agent-to-agent protocols.

await fetch(
  `${gatewayUrl}/v1/namespaces/${encodeURIComponent("acme")}/knowledge`,
  {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      Authorization: `Bearer ${token}`,
    },
    body: JSON.stringify({
      ns: "acme",
      knowledge: {
        apiVersion: "talon.impalasys.com/v1",
        kind: "Knowledge",
        metadata: {
          name: "shared-playbooks",
          namespace: "acme",
        },
        spec: {
          path: "playbooks/renewals.md",
          content:
            "# Renewal Playbook\n\nEscalate enterprise accounts when usage drops 25% week over week.",
        },
      },
    }),
    cache: "no-store",
  },
);

A complete runtime for agents

For AI-native companies operating mission critical agents in production

Durable execution

Threads persist across crashes, deploys, and cold starts so live agent state survives process turnover.

Declarative configuration

Define agent templates, MCPs, schedules, knowledge, and namespaces once. Talon syncs your manifests to production in realtime.

Bring your own cloud

Deploy it on AWS, GCP, Azure native primitives or any other cloud provider. Talon is built to scale to millions of concurrent sessions using its cloud native architecture.

Namespace isolation

Keep tools, memory, and execution boundaries scoped to the right environment so agents can operate safely in production.

Model agnostic

Works with 100+ models across the universe of providers. Works across Chat and Completions APIs.

Scheduled wakeups

Trigger agents on cron, webhooks, and product events so durable work resumes when it matters.

One debugger for the entire fleet

Sightline explorer showing namespaces, knowledge, templates, MCP servers, and runtime chat.