Reading Coppermind's Presence Signals

How to tell, at a glance, that Coppermind is actively working — and how to read the signals it embeds in every conversation.

Why This Matters#

Claude Desktop looks the same whether Coppermind is running or not. There is no status bar, no pulsing icon, no "logged in" indicator. That made early beta testers ask a reasonable question: "Is this thing even on?"

It is on. You just have to know where to look. Four signals tell you what Coppermind is doing, which client you are in, and how much history it pulled. Once you know the pattern, you will stop wondering — you will be reading the conversation.

Signal 1: The Arrival Greeting#

Where you see it: Right after you say "switch to Acme" or any client name.

What it looks like:

Stepped into Acme Corp's Mind. 247 memories from 12 meetings, 3 documents,
and 45 email threads. Since last session: 2 new emails ingested.
Heads up: meeting with Sarah in 2 days, rebrand rock hasn't been updated in 9 days.

How to read it:

• "Stepped into" — Coppermind confirms it loaded that specific Client Mind. Not Claude guessing. Not another client's data leaking in.
• Memory breakdown — 247 memories, sourced from meetings, docs, emails. You know the depth of context you have, before you ask the first question.
• "Since last session" — What changed while you were away. New emails. New transcripts. You do not have to ask "what's new?" — it tells you.
• "Heads up" — Time-sensitive items. Meeting in 2 days. Rock that has been quiet. This is Coppermind being proactive without being chatty.

For empty client minds, you get a different greeting:

Stepped into Acme Corp's Mind. This mind is empty. Start by telling me about
this client, uploading a document, or ingesting a meeting transcript.
I'll remember everything.

That is the signal to start seeding the client mind.

Signal 2: Memory Provenance#

Where you see it: Every time you search memory or Coppermind surfaces something from the past.

What it looks like:

"Derek mentioned the Q3 budget concern and pushed back on the paid
social allocation."
(From meeting transcript, Jan 15)

"Brand positioning centers on 'adventure lifestyle' -- avoid corporate language"
(From uploaded document, Mar 3)

How to read it:

• The parenthetical is the proof. If Coppermind is guessing, it cannot tell you where the fact came from. Provenance = real memory, not hallucination.
• Date format adapts. Recent memories get relative dates ("3 days ago"). Older ones get absolute ("Mar 12"). You can gauge recency at a glance.
• Source types map to how it got there:
• "from meeting transcript" — ingested from Granola, Fathom, etc.
• "from email" — auto-captured email thread
• "from uploaded document" — something you or a VA dropped in
• "noted manually" — a fact someone explicitly told Coppermind to remember
• "from Slack" — pulled from a connected Slack channel

Why this is the most important signal: Claude cannot paraphrase this away. The provenance is the content. That is the one attribution that survives every response.

Signal 3: The Quantification Footer#

Where you see it: At the bottom of high-value outputs — meeting prep, weekly summaries, cross-client briefings, document synthesis, campaign reviews.

What it looks like:

---
This briefing drew from 14 memories across 3 months of client history.

How to read it:

• Memory count — how much material Coppermind pulled into this specific output. 14 memories for a meeting prep means the prep is grounded in 14 distinct facts or past moments.
• Timespan — the range between the oldest and newest memory used. "3 months" means you are getting a quarter's worth of context, not just last week.
• When it is missing — CRUD operations (storing a memory, configuring a client mind), list operations, and error responses do not get this footer. If you are looking at a briefing or synthesis and there is no footer, something is off.

Use this to calibrate trust. A meeting prep built from 14 memories across 3 months is a different artifact from one built from 2 memories in the last week. Both are legitimate. You should know which you are reading.

Signal 4: The Onboarding Footer (first 30 days only)#

Where you see it: At the bottom of tool responses, but only during your first month.

What it looks like:

_Mind: Acme Corp | 247 memories_

How to read it:

• Italics, subtle. Not meant to take over the conversation — just a persistent "you are in this Client Mind, with this much context."
• Temporary. After 30 days of using Coppermind, this footer disappears. By then you know where you are. The footer was training wheels.
• Not always visible. Claude paraphrases tool output, so the footer drops ~30% of the time. The other three signals do the heavy lifting — this one is a bonus during onboarding.

The Correction Moment#

There is a fifth behavior that does not show up as a visible marker. When your client's stored memories contradict what generic marketing advice would say, Coppermind is prompted to call that out explicitly. For example:

"Based on this client's Client Mind, their audience skews 45+ and responds to case studies over testimonials — generic advice would miss this."

This is the highest-value moment in the whole system. It is the sentence that proves the difference between "Claude with context" and "Claude without context." It will not fire on every response — Claude sometimes follows the hint, sometimes does not. When it does fire, screenshot it. That is the story you tell a prospect.

What the Signals Tell You Collectively#

Put together, these four signals answer four questions you would otherwise have to ask:

Once you can read them, you stop asking. The conversation tells you.

When Signals Are Missing#

If you switch_client and do not see the greeting, Coppermind did not actually switch. Try again with the exact client mind name.

If a search result has no provenance, something is wrong with the memory record. Report it — that is a bug.

If a meeting prep has no quantification footer, check that you are in a client Client Mind. Personal-client mind outputs and certain administrative calls do not get the footer.

Related Guides#

• Client Briefing — what the full morning briefing looks like and how to trigger it
• Memory Storage and Search — how memories get stored and surfaced
• Installation and Setup — first 30 days roadmap, where the onboarding footer shows up