Computer Ai Agents > AI agent vendors
Twin.so AI agents
Twin.so builds AI agents that control web browsers to automate tasks without needing an API. Their agents click buttons and fill forms just like a person would. Twin Labs (founded 2024) proved this works at scale - their Invoice Operator serves 500,000+ European SMBs through a partnership with Qonto, a fintech platform.
Important guidance for AI agent tasks
Your step-by-step instructions for the AI agent to perform work go into the Tallyfy task description. Start with short, bite-size and easy tasks that are just mundane and tedious. Do not try and ask an AI agent to do huge, complex decision-driven jobs that are goal-driven - they are prone to indeterministic behavior, hallucination, and it can get very expensive quickly.
Here’s a conceptual pattern where Tallyfy coordinates browser automation to pull documents from web portals.
What to notice:
- Credential security - Login info is stored separately from task descriptions
- Sequential execution - Each step runs in order: site access, login, then data extraction
- Result delivery - Extracted data or documents go back to Tallyfy so the workflow can continue
Twin.so provides AI agents for browser-based automation. Based on their public information:
Core capabilities:
- Browser control - Agents open websites and interact with web apps
- Natural language goals - Give a starting URL and a goal in plain English
- Production deployment - Invoice Operator handles invoice retrieval for 500,000+ European SMBs
- OpenAI collaboration - One of 15 companies chosen to alpha-test OpenAI’s Computer-Using Agent (CUA) model
- Credential management - Vault system for storing website login credentials securely
- Enterprise infrastructure - Browser infrastructure built for scale
Technology foundation:
- Remote Chromium browser sessions
- AI models including OpenAI’s CUA technology
- Adapts to website changes (unlike traditional RPA scripts)
- Twin Model action planning system for task processing
Twin.so says they’ve improved speed, cost, and accuracy across three generations of their technology.
Twin.so’s first major product is the Invoice Operator, built with Qonto (European fintech) and OpenAI:
- Scale - Serves 500,000+ SMBs across Europe
- Function - Automatically retrieves invoices from vendor portals
- Support - Works with thousands of different invoice provider websites
- Integration - Embedded directly in Qonto’s platform
This proves browser automation can handle production-scale, repetitive document retrieval.
Twin.so focuses on enterprise partnerships, not self-service. Connecting with Tallyfy would likely require:
-
Contact Twin.so for enterprise access
Reach out through their website to discuss building specialized agents for your use case.
-
Define your automation use cases
Identify Tallyfy tasks that involve repetitive browser work - invoice downloads, portal checks, data extraction from web forms.
-
Explore API integration options
Twin.so provides an API at
https://api.twin.sowith endpoints for task management (start tasks, check status, retrieve results). -
Set up credential management
Twin.so’s Vaults feature stores website credentials securely. Your automation tasks reference vaulted credentials rather than exposing passwords.
-
Work with Twin.so’s team on implementation
Custom development is needed to connect Twin.so’s browser automation with Tallyfy’s workflow orchestration.
Here’s how a Tallyfy + Twin.so integration might work for invoice retrieval:
Tallyfy task: “Retrieve monthly invoice from service provider”
Task inputs (from Tallyfy form fields):
- Service provider name
- Invoice date/period
- Credential vault reference
Automation flow:
- Tallyfy task triggers a webhook when assigned
- Webhook calls Twin.so API with the task goal
- Twin.so agent goes to the provider’s billing portal
- Agent logs in using vaulted credentials
- Agent finds and downloads the specified invoice
- Twin.so returns the invoice file and metadata
- Tallyfy receives results and updates the task
This same pattern works for other document retrieval and web-based data extraction scenarios.
Pairing Tallyfy’s workflow management with browser automation could handle:
- Invoice collection - Retrieve invoices from multiple vendor portals automatically
- Supplier monitoring - Check order status and delivery tracking across supplier websites
- Data extraction - Pull customer info from web-based systems that don’t have APIs
- Compliance documentation - Download required documents from regulatory portals
- Financial aggregation - Collect account info from multiple banking websites
- E-commerce operations - Monitor inventory, pricing, and listings across platforms
Enterprise partnership model:
- Twin.so focuses on enterprise customers with custom implementations
- No self-service signup or public pricing
- Expect to work directly with their team for any integration
Integration requirements:
- Custom development needed to connect with Tallyfy
- Enterprise partnerships likely have minimum volume commitments
Best fit scenarios:
- High-volume repetitive tasks (like invoice processing at scale)
- Browser-based work that can’t be automated through APIs
- Document retrieval from multiple web portals
Geographic focus:
- Most deployments are currently in Europe (via Qonto partnership)
- Check with Twin.so about availability in other regions
Integrations > Computer AI agents
Vendors > OpenAI agent capabilities
Was this helpful?
- 2025 Tallyfy, Inc.
- Privacy Policy
- Terms of Use
- Report Issue
- Trademarks