Hosted Demo

Hosted Demo

Use the hosted demo when you want to test Dense-Mem before running Docker:

https://demo-dense-mem.markhuang.ai

The demo generates a temporary isolated team and API key. The key lasts 24 hours. Use only disposable test data.

SSO Availability

The hosted demo does not support SSO. The user portal on the demo should not show SSO login because no SSO provider is configured for that deployment.

Use the generated demo API key to access /ui, MCP, and API routes.

Related Tutorials

Goal	Tutorial
Try Dense-Mem without installing anything	This hosted demo page
Run your own local memory server	Dense-Mem Quick Start: Give Claude Code and Codex the Same Memory
Host Dense-Mem on a public HTTPS server	Secure Dense-Mem on Vultr with Traefik

What It Proves

Dense-Mem links AI instances by giving them the same external memory service.

If two clients use the same generated key, they can read and write the same temporary memory. If you generate a new key, you create a new isolated team.

This is the core Dense-Mem capability: many AI sessions can use one memory service, so every linked assistant can work with more useful context.

The AI model still reasons and responds. Dense-Mem stores evidence, claims, facts, recall results, and clarification tasks so the AI can work with more context instead of starting cold.

Limits

Limit	Demo value
Session lifetime	24 hours
Generated demo sessions	10 per IP per day
Authenticated requests	300
Write attempts	75
Save attempts	30
Content stored	128 KiB
Claims	30
Verifier attempts	10
Promoted facts	5
Recall calls	50
Rate limit	20 requests per minute

Do not store secrets, personal data, credentials, production notes, or anything critical. Expired demo teams and graph data are deleted.

Step 1: Generate A Key

Open the hosted demo page:

https://demo-dense-mem.markhuang.ai

The page creates a key automatically for the browser tab. You can also click Generate New Key.

Copy the key:

export DENSE_MEM_API_KEY="dm_..."

Use that same key in every client you want to link.

Step 2: Check MCP Tools

curl -s "https://demo-dense-mem.markhuang.ai/mcp" \
  -H "Authorization: Bearer $DENSE_MEM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

The response should include high-level tools such as:

remember
recall_memory
reflect_memories
confirm_memory

Step 3: Connect Claude Code

claude mcp add --transport http dense-mem-demo \
  https://demo-dense-mem.markhuang.ai/mcp \
  --header "Authorization: Bearer $DENSE_MEM_API_KEY"

Ask:

Use Dense-Mem to remember that this demo project codename is river-lamp and I prefer concise answers with concrete commands.

Step 4: Connect Codex

Add this to ~/.codex/config.toml:

[mcp_servers.dense_mem_demo]
url = "https://demo-dense-mem.markhuang.ai/mcp"
bearer_token_env_var = "DENSE_MEM_API_KEY"
tool_timeout_sec = 60
enabled = true

Restart Codex, then ask:

Before answering, use Dense-Mem to recall what you know about my demo project codename and response style.

Expected result: Codex should recall the memory written by Claude Code because both clients use the same demo key.

Step 5: Try A New Session

Start a new AI session with the same key configured and ask:

What does Dense-Mem remember about this demo?

The assistant should recall memory from Dense-Mem instead of depending on the old chat transcript.

Step 6: Try A Conflict

Ask:

Remember that I prefer long narrative explanations in this demo.

If the client uses the high-level memory workflow, Dense-Mem can return a clarification instead of silently replacing the earlier concise-answer preference. The assistant should ask which memory to keep.

Next

Use the demo to prove the concept. Use the local or Vultr tutorials above when you want your own server. Use Quick Start and Configuration for the manual reference.