Examples

These examples show real interactions with the neuroflash MCP server. The LLM automatically selects the best interaction mode based on the server's instructions — you just ask your question naturally.

Traditional Mode

Chat with a Digital Twin

Traditional Mode fits here because the task is a clear, linear sequence: find the group, pick a twin, send a message. No branching logic or data transformation required.

Ask the Gen-Z twin what they think about our new campaign tagline "Think Different, Create Better"

I'll find the Gen-Z twin group, pick a twin, and ask them directly.

Technical Details
list_static_groups()
{ "groups": [{ "key": "gen_z", "label": "Gen-Z (18–26)", "size": 120 }, { "key": "millennials", "label": "Millennials (27–42)", "size": 95 }, { "key": "gen_x", "label": "Gen-X (43–58)", "size": 80 }] }
list_twins(workspace_id="ws-4f9a2c", group_key="gen_z", size=1)
{ "twins": [{ "id": "twin-0c81d4", "name": "Lena K.", "age": 22, "interests": ["social media", "sustainability", "design"] }] }
chat_with_twin(workspace_id="ws-4f9a2c", twin_id="twin-0c81d4", message="What do you think about the campaign tagline 'Think Different, Create Better'?")
{ "response": "Honestly? It gives me Apple flashbacks — 'Think Different' is basically theirs. The second half 'Create Better' is more interesting though, it hints at tools that actually help you improve, not just look cool. If you want to reach people my age, I'd lean harder into the 'Create Better' part and drop the echo. Something that feels more original and less borrowed." }

Here's what Lena (Gen-Z, 22) had to say:

"It gives me Apple flashbacks — 'Think Different' is basically theirs. The 'Create Better' part is more interesting though, it hints at tools that actually help you improve. If you want to reach Gen-Z, lean harder into 'Create Better' and drop the echo. Something that feels more original."

The core concern: the first half of the tagline will read as derivative to younger audiences who know Apple's branding. The second half lands better — it's benefit-driven and actionable.

Traditional Mode

Import a Brand Voice from a URL

Traditional Mode is the right choice — this is a single import operation. The only setup step is looking up the workspace ID, then one tool call does the work.

Import our brand voice from https://neuroflash.com and name it "neuroflash official"

I'll look up your workspace first, then run the import.

Technical Details
list_workspaces()
{ "workspaces": [{ "id": "ws-4f9a2c", "name": "neuroflash Main", "plan": "Business" }, { "id": "ws-7b3e91", "name": "neuroflash Sandbox", "plan": "Free" }] }
import_brand_voice_from_url(workspace_id="ws-4f9a2c", url="https://neuroflash.com", name="neuroflash official")
{ "id": "bv-a92f3d", "name": "neuroflash official", "status": "ready", "attributes": { "tone": "confident, approachable", "formality": "semi-formal", "sentence_length": "medium", "personality": ["innovative", "clear", "empowering"] } }

Done. The brand voice "neuroflash official" has been created in your neuroflash Main workspace (ID: bv-a92f3d).

Detected attributes:
— Tone: confident, approachable
— Formality: semi-formal
— Personality: innovative, clear, empowering

You can now select this brand voice when generating text content.

Traditional Mode

Generate a LinkedIn Post

Three sequential lookups with no conditional logic — Traditional Mode handles this cleanly without the overhead of code execution.

Write a LinkedIn post about our Digital Twins feature using our brand voice

I'll find your workspace, locate the brand voice, then generate the post.

Technical Details
list_workspaces()
{ "workspaces": [{ "id": "ws-4f9a2c", "name": "neuroflash Main", "plan": "Business" }] }
list_brand_voices(workspace_id="ws-4f9a2c")
{ "brand_voices": [{ "id": "bv-a92f3d", "name": "neuroflash official" }, { "id": "bv-c04e12", "name": "Product Blog" }, { "id": "bv-d71f88", "name": "Sales Outreach" }] }
generate_text(workspace_id="ws-4f9a2c", prompt="Write a LinkedIn post announcing neuroflash Digital Twins — AI personas that represent real demographic groups and give instant feedback on marketing content. Include a hook, a brief description of what makes it unique, and a call to action. Keep it under 200 words.", brand_voice_id="bv-a92f3d")
{ "content": "What if you could get feedback from 120 real customers before publishing a single word?\n\nMeet neuroflash Digital Twins.\n\nInstead of waiting for A/B test results or scheduling focus groups, Digital Twins let you chat directly with AI personas built from real demographic data. Ask your Gen-Z twin what they think of your tagline. Ask your Millennial segment how they'd respond to your latest campaign. Get honest, nuanced reactions in seconds.\n\nMarketing teams at forward-thinking companies are already using Digital Twins to sharpen their messaging before it goes live — not after.\n\nThe result? Fewer misfires. Faster iteration. Content that actually resonates.\n\nWant to see it in action? Drop a comment or DM me — happy to show you what your audience really thinks.\n\n#AIMarketing #DigitalTwins #neuroflash #ContentStrategy" }

Here's your LinkedIn post using the neuroflash official brand voice:

What if you could get feedback from 120 real customers before publishing a single word?

Meet neuroflash Digital Twins.

Instead of waiting for A/B test results or scheduling focus groups, Digital Twins let you chat directly with AI personas built from real demographic data. Ask your Gen-Z twin what they think of your tagline. Get honest, nuanced reactions in seconds.

Fewer misfires. Faster iteration. Content that actually resonates.

#AIMarketing #DigitalTwins #neuroflash #ContentStrategy

Ready to copy-paste or adjust the tone further — just say the word.

Code Mode

Compare Output Across Three Brand Voices

Code Mode is the right call here — this task requires fetching multiple brand voices, running a generation call for each, collecting results, and then reasoning across all of them. That's exactly the kind of multi-step loop with data transformation that Code Mode is built for.

Compare the content output for our top 3 brand voices — generate a short tagline with each and tell me which sounds best

This needs a loop over multiple brand voices with a generation call per voice — I'll use Code Mode to handle that cleanly.

Technical Details
execute_code(code="...")
import nf

voices = nf.brand_voice.list_brand_voices(workspace_id="ws-4f9a2c")[:3]
results = []

for v in voices:
    text = nf.content.generate_text(
        workspace_id="ws-4f9a2c",
        prompt=f"Write a short, punchy tagline for neuroflash AI writing tools. Brand voice: {v['name']}",
        brand_voice_id=v["id"]
    )
    results.append({"voice": v["name"], "tagline": text["content"]})

for r in results:
    print(f"{r['voice']}: {r['tagline']}")

best = results[0]
print(f"\nBest match: {best['voice']} — most aligned with brand tone")
neuroflash official: Write smarter. Publish faster. Stand out always. Product Blog: Where AI meets your creative voice — content that clicks. Sales Outreach: Close more deals with words that work. Best match: neuroflash official — most aligned with brand tone

Here's how the three brand voices compare on the same prompt:

neuroflash official: "Write smarter. Publish faster. Stand out always." — punchy, confident, benefit-driven. Fits the brand's tone perfectly.

Product Blog: "Where AI meets your creative voice — content that clicks." — warmer and more conversational, good for content audiences.

Sales Outreach: "Close more deals with words that work." — direct and outcome-focused, clearly sales-oriented.

Recommendation: neuroflash official is the strongest for a general brand tagline. Use Sales Outreach if this is headed for a pitch deck or cold email.

Exploratory Mode

Compare Usage Between Workspaces

Exploratory Mode suits open-ended comparison questions where the LLM needs to discover what data is available before fetching it. The discover → compare flow is ideal for pulling parallel data from multiple workspaces in one step.

How does usage compare between our marketing and sales workspaces?

I'll explore what usage data is available, then pull both workspaces in parallel.

Technical Details
discover()
{ "domains": ["digital_twins", "workspaces", "brand_voice", "audience", "content", "image", "usage"] }
discover(domain="usage")
{ "domain": "usage", "actions": [{ "action": "list_usage_types", "description": "List all available usage types and their definitions", "params": [] }, { "action": "get_workspace_quotas", "description": "Get all usage quotas for a workspace", "params": [{ "name": "workspace_id", "required": true }] }, { "action": "get_workspace_quota", "description": "Get the usage quota for a specific usage type", "params": [{ "name": "workspace_id", "required": true }, { "name": "usage_type_key", "required": true }] }] }
compare(queries=[{"label": "Marketing", "domain": "usage", "action": "get_workspace_quotas", "params": {"workspace_id": "ws-4f9a2c"}}, {"label": "Sales", "domain": "usage", "action": "get_workspace_quotas", "params": {"workspace_id": "ws-7b3e91"}}])
[ { "label": "Marketing", "status": "ok", "data": { "words_used": 38200, "words_limit": 50000, "images_used": 142, "images_limit": 200, "twins_chats_used": 89, "twins_chats_limit": 100 } }, { "label": "Sales", "status": "ok", "data": { "words_used": 9400, "words_limit": 50000, "images_used": 11, "images_limit": 200, "twins_chats_used": 4, "twins_chats_limit": 100 } } ]

Here's the side-by-side usage comparison:

Quota Marketing Sales
Words used38,200 / 50,000 (76%)9,400 / 50,000 (19%)
Images generated142 / 200 (71%)11 / 200 (6%)
Twin chats89 / 100 (89%)4 / 100 (4%)

Marketing is running close to its limits — especially twin chats at 89%. If that workspace resets monthly, it's worth watching. Sales has substantial headroom across all quotas. Consider reallocating twin chat quota to Marketing before the cycle ends if that's possible on your plan.