limboJoin beta waitlist
Private Beta · Q1 2026

The dashboard
is dead.
Your data just
learned to talk

limbo/Chat — the first conversational music-analytics platform built as an MCP server for B2B distribution. Your catalog responds. Your artists self-serve. You scale.

Join the beta waitlist Apply to private beta
ProtocolAnthropic MCP
ClientsClaude · ChatGPT · any
Your dataStays with you
LLM costBYO credentials
BrandWhite-labelable
The problem

Five things every music
analytics dashboard
won't tell you

You don't have an analytics problem. You have a question-answering problem. The dashboard is the wrong shape for it.

PROBLEM 01

The data is there.
The question can't be asked.

Dashboards answer the questions whoever built them anticipated. The one you have right now? You'd have to wait a quarter and a Jira ticket.

PROBLEM 02

Every artist gets the
same spreadsheet.

One template for indie pop, jazz, electronica, world music. None of them are right. None of them surface what actually matters to that artist this month.

PROBLEM 03

Self-service ends at
the filter dropdown.

“What's our top territory” — that, they can do. “Which playlists drove this release vs the last one” — back to the ops team's queue.

PROBLEM 04

Reporting tools own
your data.

Most royalty / analytics platforms are run by, owned by, or funded by entities whose interests don't align with yours. Your data trains their pricing.

PROBLEM 05

Every new metric
is a new release cycle.

The world moves; the dashboard takes six months to catch up. By the time the new chart ships, the question has moved on.

The solution

One conversational
surface · three angles

The same underlying MCP server, three audiences. Each one gets a different read of the same data — but no one is hand-building new dashboards anymore.

/ ANGLE 01 · FOR YOUR LABEL

Operate the catalog,
in English

“Which releases need a Content ID claim escalated?” “Project Q3 royalties for label_2934.” “Pull all sub-clients with pre-approval queues older than 7 days.” Done. Now do five more.

Cross-tier visibilityOperational queriesForecasts in seconds
/ ANGLE 02 · FOR YOUR ARTISTS

Self-serve, finally

The artist asks the question. The MCP server answers — scoped to their data only. No SQL. No “wait for the monthly report.” No “let me circle back to ops.”

Scoped to their dataTier-3 permissioningReduces ops load
/ ANGLE 03 · FOR YOUR DATA

Sovereignty,
by design

The MCP server runs against your infrastructure. We don't ingest, don't mirror, don't train on your data. Your clients use their own LLM credentials. Zero markup. Zero exposure.

No data egressBYO LLM credentialsNot used for training
How it works

Four steps from
“installed” to answering

MCP is an open protocol. Once the server is connected, every MCP-compatible client speaks to it the same way. No custom integration per client.

01

Provision your MCP endpoint

We spin up a dedicated MCP server bound to your tenant. Your data, your scope, your credentials.

02

Connect your LLM client

Claude, ChatGPT, or any MCP-compatible client. One config line. The client points at your endpoint and authenticates with your OAuth token.

03

Define scopes per role

Admin sees everything. Level-2 client sees their sub-tree. Level-3 artist sees their own data only. Permissions enforced at the MCP layer.

04

Ask questions

That's it. The LLM client streams natural-language queries to the MCP server. The server answers in structured data the client renders as text, tables, charts.

Use cases

Four questions
worth a meeting

Four categories of conversation that currently require an analyst, a ticket, and at least one slack thread. None of them will, once the MCP server is live.

/ USE-CASE 01 · A&R

Find the next
signing

Which independent artists in Mexico crossed 100k monthly streams in the last 90 days without a label deal?Of those, which ones grew >40% MoM and have no AI-disclosed content?Pull contact info from their public socials and draft outreach.
/ USE-CASE 02 · OPERATIONS

Run the
approval queue

Show me every sub-account release pending approval > 48 hours.For label_2934, which release got rejected by QA AI most often last quarter, and why?Bulk-approve the 12 releases from artist_4408 that passed all checks.
/ USE-CASE 03 · ROYALTIES

Audit the money

Reconcile DE royalty statements between what Spotify reported and what we paid out last month.Which artists are within €500 of crossing their recoupment ceiling?Project Q3 payout per artist if streams stay flat.
/ USE-CASE 04 · ARTIST SELF-SERVE

Skip the ticket

What were my top 3 territories last month, and why?Pull my YouTube Content ID claims that haven't resolved.When does my next payout drop, and how much?
Why limbo/Chat

Four reasons it's
built this way

MCP isn't a feature. It's an architecture decision. Here's why the architecture matters more than the chat interface.

/ ARCHITECTURE

First MCP server in B2B music distribution

Anthropic's Model Context Protocol is the open standard. Building on it means your future LLM client choice isn't locked to us. Switch Claude for ChatGPT in one config line.

/ SOVEREIGNTY

Your data, your infrastructure

The server runs against your tenant. We don't ingest, don't mirror, don't train. Your clients' royalty data, your A&R pipeline, your DSP relationships — none of it leaves your scope.

/ ECONOMICS

BYO LLM credentials · zero markup

You bring your Anthropic / OpenAI API key. We don't surcharge LLM calls. You pay the LLM vendor directly at their list price. We charge for the MCP layer, not the AI tokens.

/ WHITE-LABELABLE

Branded under your domain

your-label/Chat instead of limbo/Chat. Your customers never see our brand on the conversational surface. Same depth as the rest of limbo/WhiteLabel.

Who it's for

Four shapes of partner
get it first

The Q1 2026 private beta is sized small on purpose. Four shapes of partner are the priority for first access.

/ FOR 01

White-label operators

You run a branded distribution platform on top of limbo/. Your artists need self-serve analytics. Your ops team needs to stop being a query-translation layer.

Tier-3 self-serveBranded surface
/ FOR 02

Catalog acquirers & rights holders

You've acquired a catalog and now need to audit it, value it, monitor it. Conversational queries shave weeks off the diligence cycle.

Catalog auditRoyalty forecasts
/ FOR 03

Music-tech builders

You're building a creator-facing or A&R-facing product. limbo/Chat is the AI layer underneath it. White-label the surface; bring your own LLM.

API + MCP combinedBYO LLM
/ FOR 04

Operations leaders at independent labels

You're the analyst, the comptroller, and the project manager. The MCP server compounds you. Same headcount, an order of magnitude more answered questions.

Cross-team visibilityForecasts on demand
Why now

Three things that
make this timely

The protocol, the tooling, and the market are all converging in the next two quarters. The window is small.

MCP
MCP went open

Anthropic shipped the Model Context Protocol as an open standard in late 2024. Every major LLM client now ships an MCP integration. The plumbing is done.

AI
AI literacy hit ops

Two years ago, “AI” inside a music label meant an experimental tool nobody used. In 2026, it's how the comptroller drafts the monthly statement explainer.

VC
Reporting consolidated

Major-label-aligned and VC-backed reporting tools have consolidated their position. Independents need a path that doesn't route their data through those rails.

Roadmap

Three phases.
Twelve months

Private beta in Q1, opens to white-label partners in Q2, generally available in Q3. Every phase shaped by feedback from the previous one.

/ PHASE 01 · Q1 2026

Private Beta

  • 10 hand-selected partners
  • Catalog · trends · royalties queries
  • Claude + ChatGPT MCP clients
  • OAuth 2.0 + tier-3 scoping
  • Weekly feedback loop with engineering
/ PHASE 02 · Q2 2026

White-label rollout

  • Open to all WhiteLabel partners
  • Branded under partner domain
  • Approval-queue + bulk-action ops
  • BYO LLM credentials standard
  • Custom prompt-template library
/ PHASE 03 · Q3 2026

General availability

  • Available to every limbo/ account
  • Self-serve provisioning
  • Forecasting + scenario modeling
  • Multi-modal — image + audio context
  • SDK for custom MCP-tool extensions
Legacy vs limbo/Chat

How the surface
actually changes

This isn't “the same dashboard with a chat button glued on.” It's a structurally different way of operating against the same data.

Operating surface
Legacy dashboard
limbo/Chat
Asking a new question
× Ticket the analytics team. Wait a sprint.
/ Type it. Get an answer. Iterate on it.
Artist self-serve
× Dropdown filters. Out of luck past that.
/ Full conversational access, scoped to their data.
Cross-tier audit
× Export → spreadsheet → manual pivot.
/ Asked and answered. Tier-aware out of the box.
Data sovereignty
× Your data ingested into their warehouse.
/ Server runs against your tenant. No egress.
LLM economics
× Per-seat license including AI markup.
/ BYO Anthropic / OpenAI key. Zero markup.
Brand surface
× Vendor lock-up. Their UI, their colors.
/ White-labelable under your domain.
Beta partner · name pendingPlaceholder · replaced at launch

A representative quote from a Q1 beta partner lives here. The voice will be operations-led, not founder-led — the people who actually use the MCP server daily. Replaced at launch with verbatim copy + attribution.

— Beta partner placeholder Head of operations · independent label

limbo/Chatlimbo/Analyticslimbo/Royalties
Ready

Ready for your catalog
to speak?

Q1 2026 private beta is capped at 10 partners. First-10 partners get founder-level access, custom prompt-template work, and locked-in GA pricing for the first 24 months.

Join the first 10

We'll get back within 5 working days. The application is a 4-question form — what you'd ask the system, who'd ask it, and against what catalog.

  • Founder + engineering on a weekly call
  • Custom prompt-template library, free
  • Locked GA pricing · first 24 months
  • Influence on the public roadmap
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
/ A note from Fer

We've been waiting for a
protocol like this since 2006

The reason limbo/ exists is that nobody else was going to build the infrastructure independent music needed. The reason limbo/Chat exists is the same — except now the protocol exists, the LLMs are good enough, and the market is ready.

We're not making this a product because AI is hot. We're making it because the operations people we serve have been asking the same question for fifteen years: “can the data just answer me?”

For the first time, the answer is yes. And the first ten partners on this beta get to shape what that yes looks like.

With love, since 2006.— Fer IsellaCEO & co-founder · limbo/
FI
/ CEO & CO-FOUNDERFer Isella

Bootstrapped limbo/ with Cinthia Novick in 2006. Reads the support inbox. Still answers waitlist emails personally for private beta.

20 yrs·Founder-owned
/ ENDING

Four ways to put it

Same idea, four shapes. Pick the one that lands for your team and put it on the back of a t-shirt.

The dashboard is dead.
Your data just learned to talk.
If your reports could answer back,
what would you ask first?
Built for the people
who run independent music.
Private by design.
Conversational by choice.