Answer42 MCP
An MCP tool for interactively driving the 1C:Enterprise UI through the test client: clicks, fields, tables and evidentiary scenario recording
- 6 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The tools click and fill real forms in the target base, including posting documents
- Login and password can end up in call arguments if a local credentials file is not used
Install
- Works in
- Services
Manual install
pipx install 'answer42[screenshot,linux-window-control]'The recommended PyPI install on Linux.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The tools click and fill real forms in the target base, including posting documents
- Login and password can end up in call arguments if a local credentials file is not used
This is third-party code. Review the repository files before installing.
What it does
Answer42 gives an agent UI control over 1C through the /TESTMANAGER and /TESTCLIENT pair: open a form, click a button, fill a field, pick a row from a selection form, and work with tables, dynamic lists and tabular documents. The start_session call detects the platform version from the base URL itself, brings up the infrastructure (headless via Xvfb on Linux, an interactive session on Windows), finds free ports and publishes the manager test base. A separate ui_tree feature returns a form's visible element tree with several depth and format profiles, from compact navigation to a full data dump, and large snapshots are automatically stored as an MCP resource instead of bloating the agent's context. There is a local SQLite RAG index of configuration metadata and annotated PDF-slide test recording with screenshots.
Who it is for. For teams testing 1C forms with an agent that clicks and fills fields the same way a real test-client user would.
Good fit when
- You need to run a scenario through the 1C UI rather than direct CRUD operations on the base
- You need evidentiary test-run recording with screenshots and an MCP log
- You need local configuration metadata search without an external index
Not a fit when
- You need direct base data access rather than imitating user actions in the client
- You run it as a Windows service: the client must start from an interactive user session, not as a service
Example request
Open the sales document creation form, fill in the counterparty and one item row, then show the form's element treeMCP
- Transport
- stdio
- Authentication
- not required
Limitations
This is a fork of the upstream GitLab project (platform42/answer42-mcp) with local patches on the kpsr branch, and the README does not guarantee it stays in sync with upstream. Login and password are best not passed explicitly to start_session: the values are visible to the calling agent in the arguments. On Windows the client cannot run as a service, only from an interactive session.
How to disable. Stop the Answer42 session and remove the server from your MCP client configuration.
FAQ
How does Answer42 detect the platform version?
Automatically from base_url when start_session is called, with no manual version input.
What happens to large ui_tree snapshots?
With delivery=auto, large results over the token threshold are stored as an MCP resource and handed to the agent via resource_uri instead of being inlined in the response.
Discussion
No questions or reviews yet.
Related
All 1C:Enterprise toolsAn MCP server that drives the 1C test client's interface directly: opens forms, fills fields, records and replays scenarios
An MCP server and manager for native 1C TestClient QA automation: BDD scenarios, managed-form checks, without a Vanessa Automation runtime
Tester
Тестер
A scenario testing system for 1C:Enterprise with a set of agent skills that let an AI agent operate the application like a regular user
METR
METR (MCP 1C:Enterprise Test Runner)
An MCP server that lets an agent update a 1C infobase, run YaXUnit tests and check syntax
Vanessa Automation MCP reads 1C test state from chat, showing results and the manager environment after a scenario fails.
1C AI development environment
1C AI Development Environment
1C development environment with an MCP server offering 10 tools: 11 BSL analyzers, JSON DSL to XML, dependency graph, and EPF build without 1C
README in short
The README describes three project capabilities: interactive 1C UI control, evidentiary client-test recording, and a local metadata RAG index. It details pipx installation with optional screenshot and window-control extras, the start_session call with its session parameters, and ui_tree profiles controlling depth, format and delivery of large snapshots through MCP resources.