📌 This article is for users of elvex 2.0. Find out which version you're using.
At a glance
| Context Strip | Attachment |
When | Before the first message | Any message in the thread |
Scope | Whole conversation | That and the next few messages only |
Purpose | Set the persistent foundation | Add something relevant right now |
What you pick | Space, agent, or datasource | Agent, datasource, space, artifact or new file |
Where it appears | Message input → moves to thread header | Plus icon in the message input box |
The Context Strip: Your Starting Point
Before you send your first message in a new conversation, the context strip appears in the message input box. It shows what context your conversation will begin in — your company context, personal context, a Space's context, or a specific agent or datasource.
Think of the context strip as setting the stage for the whole conversation. Whatever you select here becomes the persistent foundation for every message in that thread:
Choosing a Space means all of its pinned resources and context document flow in automatically — every response will be grounded in that team or project's shared knowledge, from the very first message.
Choosing an agent means you're chatting directly with that specialized assistant from the start.
Toggling off personal context means the conversation starts without your personal background injected.
Once you send your first message, the context strip moves up to the thread header — a persistent reminder of what your conversation is grounded in.
Key point: Context strip selections apply to the entire conversation. They're not message-level; they shape the whole thread.
Attaching a Resource: Relevant Right Now
The plus in the message input box lets you attach resources to individual messages. This is different from the context strip — you're not changing the baseline for the whole conversation. You're telling elvex: this is relevant to what I'm asking right now.
You can attach:
Agents — hand off to a specialized assistant mid-conversation
Datasources — ground elvex's response in specific internal knowledge
Spaces — bring in all the pinned resources and context from a collaborative workspace
Artifacts — pdfs, slides, dashboards, etc that you have previously created in elvex
Files — upload a document, spreadsheet, image, or other file for elvex to read and use
Attachments are flexible and composable — you can attach multiple resources at once, and elvex will read and use all of them together.
Key point: Attachments are message-level. They're for bringing in something specific and timely, not for establishing the foundation of the conversation.
When Should Add Agents to the Context Strip?
Agents are purpose-built tools configured for a specific task. Instead of asking the general elvex assistant to do something it might do okay, you can go directly to an agent built specifically for that job.
Add agents to the context strip when:
You're starting a conversation and there's an agent for your task (e.g., a Competitive Battlecard Agent, an HR Policy Agent, a Data Analyst Agent)
You want responses to adhere to a specific set of rules or instructions (e.g., brand voice, legal review criteria)
How to do it: Open Resources in the side bar and click the agent card.
When you build a new agent there is a setting that automatically adds your company and personal context to the context strip when chatting with an agent. If you would like users to chat with agent rules and context only, disable User Context in the Advanced settings.
When Should Add a Space to the Context Strip?
A Space is a shared collaborative workspace with pinned resources and shared context. When you open a conversation within a Space, all of that context flows in automatically.
Chat directly within a Space when:
You're doing work that belongs to a specific team or project and the Space already has the right agents, datasources, and context loaded
You want your conversation to benefit from everything the Space knows, without manually attaching resources
You want your conversation to impact the Space’s context (this setting has to be enabled by the Space’s owner)
How to do it: Open Resources in the side bar and click the Space's card.
When Should Add a Datasource to the Context Strip?
A datasource is a piece or set of data you have made persistent in elvex through the creation of a datasource.
Chat directly within a datasource when:
You're exploring data in an open-ended way that does not require an agent to perform a repetitive using that data
You want your entire conversation to benefit from everything the datasource knows
You want the datasource to impact your MY Space context without pinning it long term
How to do it: Open Resources in the side bar and click the datasource card.