Your Questions Answered About Itura
Find answers to common questions about deploying your agentic applications on the Itura cloud hosting platform.
Getting Started with Itura
What is Itura?
Itura is a cloud hosting platform designed specifically for deploying agentic applications. You build your AI agents using Python, and Itura handles its deployment, scaling, and monitoring.
How do I deploy my agent on Itura?
Deployment is a simple 3-step process: 1. Prepare your Python code with a Flask `/run` endpoint and a `requirements.txt` file. 2. Connect your GitHub repository containing the code. 3. Add any necessary environment variables (like API keys) via the Itura UI. Itura automatically builds, deploys, and provides a secure endpoint.
What code preparation is needed?
You need a Python script with a Flask web server defining a `/run` POST endpoint that triggers your agent logic. You also need a `requirements.txt` file listing all Python dependencies for your agent.
How does GitHub integration work?
Connect your GitHub repository to Itura. When you push new commits to your repository, Itura automatically detects the changes, rebuilds your agent container, and deploys the new version, ensuring your endpoint always runs the latest code.
How do I provide secrets like API keys to my agent?
You can securely add environment variables through the Itura web UI. These variables are injected into your agent's runtime environment, making them accessible to your code without exposing them directly.
How do I invoke my deployed agent?
Once deployed, Itura provides a unique URL (e.g., `https://your-agent-name.agent.itura.ai`) and a secret API key. You invoke your agent by sending an HTTP POST request to the `/run` endpoint of this URL, including the API key in the `Authorization: Bearer <key>` header and your input data in the JSON body. The agent runs asynchronously; you get a `run_id` back and can check status via the UI or a `/status/<run_id>` endpoint.
Technical Details & Specifications
What languages and frameworks are supported?
Itura supports Python (versions 3.9, 3.10, 3.11, 3.12, 3.13). You can use any Python library, any framework (like Langchain or CrewAI), and any LLM. As long as your agent is triggered via a Flask `/run` endpoint and dependencies are listed in `requirements.txt`, it will work on Itura.
What infrastructure does Itura run on?
Itura is a serverless platform built on AWS and Google Cloud Platform, using Docker containers to run your agent. You don't need to manage any servers or infrastructure.
What are the resource limits per agent invocation?
Each agent invocation gets 1GB of RAM and 10GB of temporary disk storage.
What is the maximum execution time for an agent?
An agent invocation can run for a maximum of 10 minutes.
Is there a limit on the size of my agent's code?
Yes, the total size of your code package (including dependencies) is limited to 1GB per agent.
How does logging work?
Any output your agent prints to standard output (stdout) or standard error (stderr) is automatically captured. You can view these logs in the Itura UI. Log retention duration depends on your subscription plan.
Security
How does Itura secure my agent and data?
Itura runs each agent in a secure, isolated containerized environment ensuring that your agent's code and data are not exposed to other users. All communication between Itura and your execution environment is secured using SSL/TLS. Environment variables secrets are encrypted at rest and are injected securely to your agent at runtime. Itura does not store your execution data apart from the logs and print statements generated by your agent.
Pricing & Billing
How does Itura pricing work?
Itura uses a tiered subscription model (Free, Plus, Pro) based on usage limits like monthly invocations, total execution time, concurrent runs, and log retention. You pay a predictable monthly fee for the hosting infrastructure. See our Pricing page for details.
What counts as an 'invocation'?
An invocation is counted each time your agent's `/run` endpoint is successfully called via its unique Itura URL.
How is 'total execution time' calculated?
This is the sum of the compute duration (measured in 1-second increments) for all your agent invocations within a billing month. Each invocation's time is capped at the 10-minute maximum.
What happens if I exceed my plan's limits?
If you exceed the monthly invocation or execution time limits of your plan, your agent endpoints will temporarily become unavailable until the start of the next billing cycle. We recommend upgrading your plan if you anticipate higher usage.
Are costs for external APIs (like OpenAI/Anthropic) included?
No. Itura's pricing covers the agent hosting infrastructure (compute, scaling, security, logging). You are responsible for managing and paying for any external services your agent uses, such as LLM APIs. These costs are typically handled via API keys passed as environment variables.
Can I change my subscription plan?
Yes, you can upgrade or downgrade your plan at any time through the Itura dashboard. Changes typically take effect at the beginning of your next billing cycle.
Platform & Features
Who is Itura designed for?
Itura is built for AI/ML developers, engineers, startups, and Python developers who want to quickly deploy agentic applications to the cloud without managing infrastructure.
What kind of agents are suitable for Itura?
Itura is ideal for agents performing specific tasks like automation (data entry, workflows), content generation (drafting, summarizing), and information gathering (web scraping, research).
Does Itura support automatic updates?
Yes, through its GitHub integration. Every push to your connected repository triggers a new deployment, making your latest agent version live automatically.
What new features are planned?
We are continuously improving Itura. Upcoming features include support for custom domains, scheduled agent execution, enhanced monitoring and alerting, and options for storing output data.