AI Tool Finder for Files
In-browser · data sent outDescribe a file task in plain language and get the right tool with parameters or steps. Free, no sign-up — bring your own LLM key, it never touches Vecaria.
Runs in your browser · Routes across the full on-site tool catalog · Your key never touches us
Two data paths: with your own LLM key, requests go directly from your browser to the endpoint you enter — neither the key nor the content passes through Vecaria; without a key, the query (carrying no key) is sent to our semantic-search endpoint to match tools. Neither path stores your input.
Plans (including multi-step sequences) are generated in one shot — nothing is executed step by step and plans cannot be edited or re-run in place; open a step's tool page to confirm parameters and run.
What is it? AI Tool Assistant turns a natural-language description of a task into a concrete plan: which tool to use, with which parameters, or a sequence of steps. It only recommends tools from Vecaria's catalog of browser-based file tools, never external services. You can bring your own LLM key, in which case requests go directly from your browser to your endpoint and never touch Vecaria.
TL;DR: Describe a task in one sentence and get the right tool with parameters — free, no sign-up, bring your own LLM key.
How to use
- Describe the outcome you want in one sentence — e.g. "Convert STEP to STL and simplify the mesh". Tool names are not required.
- Optionally enter your LLM base URL, model and API key.
- Click Plan to get tool suggestions or a step sequence.
FAQ
Is my API key stored?
Only if you tick "Remember in this browser" — otherwise it stays in memory for the session.
Which model is used by default?
The default model is configured site-wide (admins can update it without a redeploy); you can always override the model and base URL under "Advanced". If no site default is set, a built-in fallback is used.
What happens without a key?
The assistant falls back to Vecaria's semantic search: your query text is sent to our own VPS for embedding-based matching — nothing is stored, and there is no Pro cloud path.
Where do the suggestions come from?
The assistant reads this site's public tool catalog (names, categories, aliases) and only recommends tools from it. Plans are generated by the LLM you configure — verify tool and parameters before use.
Which tool should I use?
Describe your task in one sentence and this assistant will suggest the tool and parameters; or use semantic search to find tools by meaning.
What are the ways to find a tool on this site?
Three ways: keyword search & categories (fastest when you know the name); semantic search (describe the meaning); the AI tool assistant (describe the task, get tool + parameters — this page).
How is this different from the search bar?
Keyword search matches tool names and categories — fastest when you already know what a tool is called. The assistant reads a whole task sentence and returns the tool plus its parameters, or an ordered step sequence, so it works when you don't know the name.
Which tools can it recommend?
Only tools from Vecaria's own catalog — browser-based CAD and 3D, GIS, PCB, CNC, PDF and document utilities. It never recommends external services or paid third-party products.
Do I need an account or a subscription?
No. The assistant needs no sign-up and works within the free daily allowance. For stronger planning, bring your own LLM API key — it is stored only in your browser and never sent to Vecaria.
Can it plan multi-step tasks?
Yes — describe the whole job, e.g. "turn a scanned drawing into a clean DXF", and it returns an ordered step sequence. Plans are one-shot suggestions: open each step's tool page to confirm the parameters before you run it.
Examples
- Convert STEP to STL and simplify the mesh
- Merge several DXF drawings into one file
- What is the board thickness in this Gerber?
- Turn a scanned PDF into searchable text
- Which tool can view GeoJSON?
Limitations
- With a key set, your request content is sent to the third-party LLM service you configure — it leaves your device, so don't submit sensitive or confidential information.
- Without a key, only semantic tool matching is available (no step planning); the query is sent to Vecaria's own VPS and not stored.
- Plan quality depends on the LLM you provide.
- Plans (including multi-step sequences) are generated in one shot: nothing is executed step by step and plans cannot be edited or re-run in place — open a step's tool page to confirm parameters and run.
Privacy
The core processing happens outside your browser: with an LLM key set, your request goes directly from your browser to the third-party LLM service you configure — the content leaves your device and is governed by that provider's privacy policy, so don't submit sensitive or confidential information. Without a key, the query text is sent to our own semantic-search endpoint (our VPS, nothing stored). Your key never touches us, and neither path stores your input.