Your AI's private notebook
Context gives your AI assistants what they normally lack: access to your documents' content and a persistent memory that spans conversations and even different AI agents. You control both from a single dashboard.
Context is an AI-agent-agnostic context layer that works seamlessly with Claude, Cursor, ChatGPT, and any MCP-compatible tool, enabling all your AI assistants to share the same knowledge and memory.
What is Context?
AI assistants like Claude, Cursor, and ChatGPT are powerful, but every new conversation starts with little to no prior context. With Context, your assistant always has access to your documents and the memories and notes saved from previous conversations, enabling more accurate, personalized, and consistent responses.
Context is a private knowledge space where your assistant can retrieve information and store new memories. Only you and your AI agents hold the key to that knowledge.
You upload documents (notes, manuals, reports). Your assistant can then search them and answer questions using their content, instead of guessing.
When your assistant learns something worth keeping (your preferences, decisions, facts about your work) it saves a short note. Next conversation, it remembers.
How do I use it?
Three steps, no coding needed. The dashboard walks you through each one.
https://context.veeliq.com/mcpSome clients only accept a plain URL and cannot sign you in. For those, the Connection section gives you a personal URL with a token in it. Treat that token like a password: whoever holds it reaches your Context. Prefer signing in above whenever your assistant supports it.
What your assistant can do
Once connected, your assistant gains the 33 tools below. It picks the right one on its own, so you never call them by name. They are listed here so you can see exactly what a connected assistant is able to reach and change.
search_knowledgereadFinds the passages of your documents that answer a question.
list_sourcesreadLists what you have uploaded, its folder, and whether it finished processing.
list_source_foldersreadLists the folders your documents are organised into, so a search can be narrowed to one.
get_sourcereadOpens one document in full instead of just the matching pieces.
get_pending_pagesreadFetches the pages of a scanned document that nobody can search yet.
submit_transcriptionwriteSaves the text your assistant read off one of those scanned pages.
rememberwriteSaves a fact you asked it to keep. If it contradicts something already saved, it stops and asks instead of guessing.
recallreadBrings back the saved facts relevant to what you are talking about.
forgetdeleteDeletes one memory for good.
list_instructionsreadShows the index of procedures you have saved.
search_instructionsreadFinds the procedure that fits what you are trying to do.
get_instructionsreadLoads one procedure in full before following it.
save_instructionswriteWrites down a procedure so it can be reused later. Only with your say-so.
delete_instructionsdeleteRemoves a saved procedure for good.
list_notesreadSees the names of your notes and their folders, without opening them.
list_note_foldersreadSees how you have organised your notes into folders.
get_notereadOpens one note in full when it is relevant to what you asked.
search_notesreadFinds the note where you wrote about something.
save_notewriteWrites a note for you, when you ask it to. It cannot create folders.
delete_notedeleteRemoves one note for good, after asking you to confirm.
list_tasksreadShows the task list, most pressing first.
get_taskreadOpens one task with its full history.
search_tasksreadChecks whether something was already proposed, approved or turned down.
propose_taskwriteSuggests new work, with the evidence for it. Nothing starts until you approve.
update_taskwriteEdits a task or moves it along.
decide_taskwriteYour call on a proposal. Reserved for you, not for the assistant.
graduate_taskwriteAttaches a saved procedure to approved work and starts it.
get_context_briefreadOne summary of your space so the assistant knows where it is.
get_by_idreadOpens the exact item behind an id you copied from a card here.
switch_workspacewriteMoves between your spaces, for example personal and work.
list_stalereadPoints out what has gone long enough without review to be worth a look.
verifywriteMarks one item as still true, resetting its clock.
verify_manywriteSame, for a whole list at once.
Is my data private?
- Your space is yours alone. Every document and memory is tied to your account and invisible to anyone else.
- If you connected by signing in, you can withdraw your assistant’s access whenever you want. If you used the URL with a token instead, treat that token like a password: revoke it in one click and create a new one.
- You can delete any document or memory at any time, permanently.
Common questions
The AI you already use: Claude, Cursor, or ChatGPT. Context does not replace it; it gives it a better memory.
The plug that connects your assistant to Context. You will see the letters in the setup, but you never need to touch it beyond copy and paste.
The free plan lets you try everything with one document and a small memory. Pro raises the limits.
Claude (app and Claude Code), Cursor, and ChatGPT. Anything that speaks MCP can connect.