OmniDimension Python SDK
Build and ship AI voice agents from a single prompt.
OmniDimension lets you build, test, and deploy reliable voice AI assistants by simply describing them in plain text. The platform offers rigorous simulation testing and real-time observability, making it easy to debug and monitor agents in production.
👉 Try the Web UI — You can also build and test voice agents visually using our no-code interface.
🚀 Features
- Prompt-based creation: Define voice agents with natural language.
- Drag-and-drop editor: Chat or visually edit flows, voices, models, and more.
- Prebuilt templates: Use plug-and-play agent templates for common use cases.
- Testing & monitoring: Simulate edge cases and debug live calls.
- Knowledge Base support: Upload and attach documents (PDFs) for factual grounding.
- Integrations: Connect to external APIs, CRMs, or tools like Cal.com.
- Phone agents: Assign numbers and initiate real voice calls via the SDK.
- Bulk Call: Send multiple voice calls AI to multiple numbers simultaneously.
📦 Installation
Basic SDK
pip install omnidimension
Requires Python 3.9+
🔐 Authentication
First, obtain your API key from the OmniDimension dashboard. Store it in your environment variables:
Linux/macOS
export OMNIDIM_API_KEY="your_api_key_here"
Windows (CMD)
set OMNIDIM_API_KEY=your_api_key_here
In Python
import os
from omnidimension import Client
api_key = os.environ.get("OMNIDIM_API_KEY")
client = Client(api_key)
✨ SDK Usage
from omnidimension import Client
# Initialize the client with your API key
client = Client(api_key="your_api_key")
# List agents
agents = client.agent.list()
print(agents)
🛰️ MCP server
The OmniDimension MCP server lives in its own repositories, not in this SDK. Pick whichever fits your client.
Hosted (no install)
Point any MCP client at the hosted server. It speaks the standard OAuth flow, so clients discover the rest automatically:
https://mcp.omnidim.io/mcp
For Claude Code:
claude mcp add --transport http omnidim https://mcp.omnidim.io/mcp --scope user
Source: Omnidim/omnidim-mcp-cloud
Local (stdio)
For clients that launch a local server (Claude Desktop, Cursor, Windsurf), use the npm package @omnidim-ai/mcp-server. Save this as your MCP client config:
{
"mcpServers": {
"omnidim": {
"command": "npx",
"args": ["-y", "@omnidim-ai/mcp-server"],
"env": {
"OMNIDIM_API_KEY": "<your_omnidim_api_key>"
}
}
}
}
Source: Omnidim/omnidim-mcp-server
📡 Providers API
The Providers API allows you to discover and explore all available AI providers for LLMs, voices, STT, and TTS services.
Quick Examples
# Get all LLM providers
llms = client.providers.list_llms()
print(f"[SUCCESS] Found {llms['total']} LLM providers")
# List voices with pagination
voices = client.providers.list_voices(page=1, page_size=50)
print(f"[SUCCESS] Found {voices['total']} voices")
![INFO] Complete Providers Documentation
📚 Knowledge Base
files = client.knowledge_base.list()
print(files)
file_ids = [123]
agent_id = 456
response = client.knowledge_base.attach(file_ids, agent_id)
print(response)
🔌 Integrations
response = client.integrations.create_custom_api_integration(
name="WeatherAPI",
url="https://api.example.com/weather",
method="GET"
)
print(response)
client.integrations.add_integration_to_agent(agent_id=123, integration_id=789)
☎️ Phone Number Management
numbers = client.phone_number.list(page=1, page_size=10)
print(numbers)
client.phone_number.attach(phone_number_id=321, agent_id=123)
# Search for numbers available to purchase in a region
# carrier is region-specific; every region currently requires one. Omitting it
# (or passing an unknown value) gets a 409/404 back listing the valid carriers.
available = client.phone_number.search(region="US", pattern="415", page=1, limit=20, carrier="carrier-us")
print(available)
# Purchase one of the numbers found above
purchase = client.phone_number.purchase(
region="US",
phone_number="+14155550123",
carrier="carrier-us",
idempotency_key="order-2024-12-01-001" # safe to retry with the same key
)
print(purchase)
# Release a number you no longer need
client.phone_number.release(phone_number="+14155550123")
Reseller accounts can pass user_id to list, search, purchase, and release to act on a
specific child client's account instead of their own.
🤝 Reseller Management
Reseller-only methods for managing child organizations, users, credits, and KYC.
# List child organizations
orgs = client.reseller.list_organizations()
# Create a new child user
new_user = client.reseller.add_user(
name="Jane Doe",
email="jane@example.com",
phone="+14155550123",
password="a-strong-password",
welcome_minutes_to_credit=100,
cost_per_min=0.05
)
# Control which dashboard menus a child user can see
client.reseller.set_access_control(
user_id=456,
dashboard_menu_access={"billing": True, "agents": True, "integrations": False}
)
# Set or clear a child account's expiry date
client.reseller.set_expiry(user_id=456, expiry_date="2025-12-31")
client.reseller.set_expiry(user_id=456) # clears the expiry
# Set a child organization's concurrent call limit (absolute value, not a delta)
client.reseller.set_concurrency(child_organization_id=789, new_limit=10)
# Calculate and transfer credits
cost = client.reseller.calculate_credits(minutes=500, cost_per_min=0.05)
client.reseller.transfer_credits(to_organization_id=789, minutes=500, cost_per_min=0.05)
# Revert a previous transfer (uses the original rate, no cost_per_min to pass)
client.reseller.revert_credits(from_organization_id=789, minutes=500)
# Credit transfer/revert history
logs = client.reseller.credit_logs(page=1, page_size=30)
# KYC: check status, then walk next_step until it reports "completed"
# carrier decides both the requirements and the step flow; every region
# currently requires one, and step names are only knowable from this call.
status = client.reseller.kyc_status(user_id=456)
requirements = client.reseller.kyc_requirements(region="IN", carrier="carrier-1")
client.reseller.submit_kyc_step(
step="register", user_id=456, region="IN", carrier="carrier-1", full_name="Jane Doe")
📞 Bulk Call Management
# First, get your phone number IDs
phone_numbers = client.phone_number.list()
phone_number_id = phone_numbers['phone_numbers'][0]['id'] # Use the first available phone number
# Create a bulk call campaign
contact_list = [
{
"phone_number": "+1234567890",
"customer_name": "John Doe",
"product_interest": "Premium Plan"
},
{
"phone_number": "+1987654321",
"customer_name": "Jane Smith",
"product_interest": "Basic Plan"
}
]
# Create immediate bulk call
bulk_call = client.bulk_call.create_bulk_calls(
name="Marketing Campaign - Q4 2024",
contact_list=contact_list,
phone_number_id=phone_number_id
)
# Create scheduled bulk call
scheduled_call = client.bulk_call.create_bulk_calls(
name="Scheduled Campaign",
contact_list=contact_list,
phone_number_id=phone_number_id,
is_scheduled=True,
scheduled_datetime="2024-12-25 10:00:00",
timezone="America/New_York"
)
# Create a draft campaign with number rotation and call conditions
draft = client.bulk_call.create_bulk_calls(
name="Draft Campaign",
contact_list=contact_list,
phone_number_id=phone_number_id,
save_as_draft=True,
call_conditions=[
{"column": "product_interest", "operator": "=", "value": "Premium Plan"}
],
rotation={
"numbers": [
{"phone_number_id": phone_number_id, "sequence": 1},
{"phone_number_id": 456, "sequence": 2}
],
"strategy": "round_robin",
"calls_per_number": 50
},
concurrent_call_limit=5
)
bulk_call_id = draft['json']['id']
# Add contacts to an existing campaign
client.bulk_call.add_contact(
bulk_call_id=bulk_call_id,
to_number="+1234567890",
custom_variables={"customer_name": "John Doe"}
)
client.bulk_call.add_contacts(
bulk_call_id=bulk_call_id,
contacts=[
{"to_number": "+1987654321", "custom_variables": {"customer_name": "Jane Smith"}}
] # up to 1000 contacts per request
)
# Start the draft campaign
client.bulk_call.start_bulk_call(bulk_call_id=bulk_call_id)
# Fetch all bulk calls
bulk_calls = client.bulk_call.fetch_bulk_calls(page=1, page_size=10, status="active")
# Get bulk call details
details = client.bulk_call.detail_bulk_calls(bulk_call_id=123)
# Control bulk call actions
client.bulk_call.bulk_calls_actions(bulk_call_id=123, action="pause")
client.bulk_call.bulk_calls_actions(bulk_call_id=123, action="resume")
client.bulk_call.bulk_calls_actions(
bulk_call_id=123,
action="reschedule",
new_scheduled_datetime="2024-12-26 14:00:00",
new_timezone="America/New_York"
)
# Cancel bulk call
client.bulk_call.cancel_bulk_calls(bulk_call_id=123)
# Live status of a running campaign
status = client.bulk_call.get_live_status(bulk_call_id=123)
# Call lines with cursor pagination and filters
lines = client.bulk_call.get_bulk_call_lines(
bulk_call_id=123,
pagesize=50, # max 150
call_status="completed",
include_total=True
)
next_page = client.bulk_call.get_bulk_call_lines(
bulk_call_id=123,
cursor=lines['json']['next_cursor']
)
# Retry failed calls manually
client.bulk_call.manual_retry(bulk_call_id=123)
# Change the concurrent call limit of a campaign
client.bulk_call.update_concurrency(bulk_call_id=123, concurrent_call_limit=10)
# Manage the campaign's number rotation
numbers = client.bulk_call.list_rotation_numbers(bulk_call_id=123)
client.bulk_call.add_rotation_number(bulk_call_id=123, phone_number_id=456)
client.bulk_call.set_rotation_number_active(bulk_call_id=123, assignment_id=789, is_active=False)
📁 Recommended Project Structure
/docs/
├── agents/
├── calling/
├── integrations/
├── knowledge_base/
└── phone_numbers/
/examples/ # Sample Python scripts
/cookbook/ # Ready-made project use cases
🌐 Learn More
Visit omnidim.io to explore the full platform, UI builder, and templates.
💬 Support
- Ask in our Discord, where engineers from the team answer build questions.
- Bugs and feature requests: open an issue.
- Account or billing: support@omnidim.io.
Release files for omnidimension 0.4.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| omnidimension-0.4.1.tar.gz | 31.9 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| omnidimension-0.4.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 66.6 kB
Release files / omnidimension-0.4.1.tar.gz
| Download URL | omnidimension-0.4.1.tar.gz |
|---|---|
| Size | 31.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
233981a6e1e7d1e8e6e78621656e76f3a904a5240497e909897ce3c90bbe11c4
|
|
BLAKE2b-256 checksum How to use checksums |
2ec49995f9d69cfc923386180f0c820c9ffec7677a2eefd4adcd19d22fb0aab3
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 8, 2026.
Transparency logRelease files / omnidimension-0.4.1-py3-none-any.whl
| Download URL | omnidimension-0.4.1-py3-none-any.whl |
|---|---|
| Size | 34.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
ad80494c2e08437376151722f4914a113fd7f22009a069537960ecfed1e83670
|
|
BLAKE2b-256 checksum How to use checksums |
717c5f1108f5e9f750ce1beacb9ebe35211c25b3eef6e56c4e0a52166cecb202
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 8, 2026.
Transparency log