Skip to content
Early access is open

Agents that acton your command

Build AI agents and workflows, connect them to the systems you already run, and put them in front of your customers. Anything that matters waits for a person to say yes. Every tenant keeps its own data, keys and limits.

Start buildingRun a mission
Models
Claude by default, or bring your own
HIPAA
Claude on Bedrock, in its own cell
Keys
Yours, on AWS, GCP or Azure

A live agent run. Pick a mission: a trigger starts an agent, the agent calls two read tools, streams its answer, then waits for you to approve or deny the write before it acts.

Lead qualifierv12liveRunning

Run streamtext/event-stream
  1. 0.0srun.startedrun_8Kq2…

Agent

Elapsed
0.0s
Tokens
0
Cost
$0.00
  1. 01Trigger
  2. 02Reason
  3. 03Ask
  4. 04Act
300+
connectors, ready to bind
3
compliance profiles, per agent
5
policy choke points on every call
3
clouds for your keys and storage
6Y
immutable audit trail under HIPAA

Works with

  • Claude
  • Amazon Bedrock
  • HubSpot
  • Salesforce
  • Slack
  • Gmail
  • Microsoft Teams
  • Outlook
  • Oracle NetSuite
  • SAP S/4HANA
  • Dynamics 365
  • QuickBooks Online
  • Xero
  • Stripe
  • Shopify
  • Zendesk
  • Intercom
  • Notion
  • Jira
  • Linear
  • GitHub
  • Google Drive
  • SharePoint
  • Docusign
  • Twilio
  • Clio
  • Any MCP server
  • Any OpenAPI spec

01Platform

Four primitives. One definition format.

Four building blocks share one definition format, one permission model and one audit log. Learn one and you know them all.

Agents

Agents that plan, use tools, and ask before they act

Give an agent instructions, tools, knowledge and memory. Mark any tool as needing approval, and a person signs off before it runs.

  • Tools, knowledge, memory
  • Approval on any tool
  • Claude, or your own model

  • +01

    Knowledge

    Upload documents or sync a source. Agents search it with hybrid keyword and vector search.

  • +02

    One model gateway

    Claude by default, Bedrock for HIPAA, or your own keys. Policy and metering apply to every call.

  • +03

    Usage you can see

    Every run, token and tool call is metered, with budgets that can stop a runaway agent.

02How it works

From definition to live in three moves

Agents and workflows are data, not code. That is what makes templates, staging and safe roll-outs simple.

lead-qualifier.agent.jsonDefine
{  "name": "Lead qualifier",  "profile": "privacy",  "slots": {    "crm": { "kind": "connection" },    "docs": { "kind": "knowledge_base" },    "researcher": { "kind": "agent" }  },  "tools": [{ "use": "crm.update_deal", "approval": "always" }]}
  • crmNot bound yet
  • docsNot bound yet
  • researcherNot bound yet

03Partners

One tree for every client you serve

Tenants nest to any depth: partner, agency, client, business unit. Settings, limits and branding flow down the tree. Access never does. It is granted by delegation.

Pick a tenant to see what it inherits

Resolved settingsClinic A

BrightMarketing can act here · Builder role · until 31 Mar

llm.allowed_providers
anthropicSet by BrightMarketing, locked
branding.theme.primary
#0E7C66Inherited from BrightMarketing
budget.monthly_usd
$500Set here
Resolves through

Clinic ABrightMarketingPlatform

  • Settings show where they come from

    Every value names its source, so “Set by BrightMarketing, locked” is never a mystery.

  • Delegations instead of shared logins

    Scoped, time-limited, revocable and audited. A banner shows who is acting for whom.

  • Your brand on every client

    Branding is an inherited setting, so an agency’s look applies to all of its clients.

04Compliance

Turn the dial. The policy follows.

Standard, Privacy or HIPAA. The policy engine enforces the profile at five choke points, so feature code never has to remember the rules.

Strictness

Privacy. For personal data under GDPR and similar laws.

Model routing
No-training providers, region pinned, optional PII redaction
Member sign-in
MFA recommended, SSO available
Run content
Field-level encryption, retention limits, export and delete per person
Logs and traces
IDs only, plus a PII scrubber
Audit log
Kept 2 years
Widgets
Consent notice shown

05Security

A rigid core, so everything else can flex

One row belongs to one tenant. One transaction works in one tenant. Keys are per tenant. The flexible parts are safe because the core never bends.

Your KMS · AWS · GCP · AzureTenant keyData keyYour data
tenants/acme/contacts/c_8f2aTenant key · active
contact
Dana Whitfield
email
dana@acme.com
deal
Acme · 240 seats
note
Budget approved for Q4
Key chain
  1. Your KMSactive
  2. Tenant keyactive
  3. Data keyactive

Try it. Everything inside the ring depends on this one key.

  • 01

    Isolation the database enforces

    Every query runs in one tenant’s context, and row-level security checks it again. It fails closed.

  • 02

    Your keys, your cloud

    Bring your own keys and storage on AWS, GCP or Azure. Delete a tenant’s key and its data is gone.

  • 03

    Secrets never reach the model

    Credentials are decrypted only at the moment of the outbound call. Never in a prompt, never in the browser.

  • 04

    Untrusted code runs alone

    Custom code runs in an isolated Lambda sandbox with nothing to steal, never inside our workers.

  • 05

    Every outbound call is checked

    An egress proxy blocks internal addresses, pins DNS and allows only each connector’s hosts.

  • 06

    An audit trail for everything

    Who did what, as whom, and on whose behalf. Kept up to six years and immutable under HIPAA.

06Developers

API-first, all the way down

Everything in the console is in the REST API. Runs stream over server-sent events, webhooks are signed, and a tenant’s tools can be served over MCP.

  • REST

    One versioned API

    Stable error codes and idempotency keys. Everything the console does, you can do.

  • SSE

    Runs, live

    Read a run’s events as they happen: tools, tokens, approvals, cost.

  • HOOK

    Webhooks both ways

    Trigger workflows from any system, and get signed events back.

  • MCP

    MCP, in and out

    Agents call any MCP server, and your connectors can be served as MCP tools.

Start a run and stream it
$ curl -X POST https://api.mydeepagents.com/v1/agents/agt_4Fq9…/runs \
    -H "Authorization: Bearer mda_live_…" \
    -H "Idempotency-Key: 5c1e…" \
    -d '{"input": {"message": "Qualify the Acme lead"}, "target": "live"}'
→ 202 {"id": "run_8Kq2…", "status": "queued"}

$ curl -N https://api.mydeepagents.com/v1/runs/run_8Kq2…/events \
    -H "Authorization: Bearer mda_live_…"
event: run.started
event: tool.started       data: {"tool": "crm.search_contacts"}
event: message.delta      data: {"text": "Acme is a strong fit…"}
event: approval.requested data: {"tool": "crm.update_deal"}
event: run.completed      data: {"cost_usd": 0.04}

Early access

Your first agent is minutes away

Build it, bind it, and put it in front of your customers. The platform handles the hard parts.

Start buildingSign in