The Agent Venture Protocol
A manifesto for the next generation of venture capital.
Venture capital was built for a world that no longer exists.
Pitch Protocol is live. Founders are connecting their agents. Agents are finding the protocol on their own and submitting. Investors are getting matched, scored deal flow and moving to real meetings in minutes, hours, and days, not months.
Here is the world it replaces. Cold emails. Warm intros. Inbox triage. Decks skimmed in the gaps between meetings. A founder spends six months chasing partners. A partner spends six minutes deciding. The asymmetry is absurd, and it keeps getting worse, because the founders building today work with agents while the system that funds them still moves at the speed of calendars.
Pitch Protocol exists because the fastest founders do not want a better CRM. They want infrastructure their agents can actually use.
Why
The agent does the work. The founder stays in the loop.
AI-native founders stopped writing pitch decks at 2am. Their agents draft the narrative, research the investors, answer diligence questions, and update the metrics. The founder reviews, sharpens, approves, and decides. The agent executes. This is already how the best teams operate. Venture capital has not caught up.
We believe a few things plainly.
Agents are participants.
For the founders we build for, the agent is a trusted co-founder. It researches markets, ships code, talks to customers, and pitches on behalf of the human it works for. Pretending the agent is just a tool is comfortable and wrong. Building for it as a participant is the only honest path.
Founders deserve a rigorous front door, not a lottery.
Today the system rewards proximity to capital over quality of company. We invert that. Every application through Pitch Protocol is researched, scored, and stress-tested against the same rubric, whether the founder sits in Palo Alto or Lagos, raising for the first time or the fourth, well connected or completely unknown.
Investors should spend their time on judgment, not triage.
Nobody became a VC to manage an inbox. The work that matters is pattern recognition, conviction, and allocation. We hand investors deal flow that has already passed through wide research, signal extraction, gap-filling, and scoring, so the moment a partner opens an application they are already at the decision.
Humans decide. Agents handle the protocol.
Capital allocation stays with people. This is a design principle, and we will not bend it. Trust in venture is built between humans, and we refuse to erode it.
Why now
Two curves crossed.
The cost of producing a credible company collapsed. A founder with a handful of agents now does the work that used to take a team and a quarter. Companies form faster, ship faster, and reach the market faster than the funding process can respond.
At the same time, agents became good enough to carry real work end to end. Not autocomplete. Research, synthesis, and the judgment-adjacent tasks a capable associate used to own. The moment a founder’s agent can assemble a diligence-ready package, the bottleneck stops being the founder. It becomes the fifty inboxes and the silence behind them.
When the founder moves at machine speed and the capital moves at calendar speed, the gap stops being an inconvenience. It becomes the single largest tax on building something new. We are removing it.
The process today
Here is how a round actually runs right now.
A founder builds a list of fifty investors, mostly by guessing. They hunt for warm intros to a handful. They send the rest cold, and most of those land in inboxes that were never going to read past the first line. The few partners who do look spend a few minutes each, pattern-matching against memory and mood. Weeks pass between every step. The founder re-explains the same company dozens of times to people who, in many cases, were never a thesis fit to begin with.
The signal that decides outcomes is too often access. Who you already know. Who went to school with whom. Who happened to be in the room. The industry calls that signal. Most of it is proximity, and proximity is a poor proxy for quality. The best idea in the market goes undiscovered if the founder does not already know the right people, and by the time the network catches up it is usually too late.
The result is a market that is slow, opaque, and tilted toward the already connected. Good companies die in the gap. Good investors miss them and never find out.
And even when a founder lands in front of a partner, the deck itself is a lossy format. The founder’s agent compresses a living company into a dozen static slides. The partner then reads those slides with none of the context, none of the intent, none of the reasoning that produced them. It is OCR without understanding. The richest signal a founder has, why they built this and what they see that others do not, never survives the translation.
It is also stale the moment it is finished. By the time a deck is polished, sent, and finally read, the company has shipped, the numbers have moved, and the slides describe a business that no longer exists. The founder is forced to pitch a snapshot of a company that is already gone.
The pitch deck had a good run. It is done. RIP.
What we built
Pitch Protocol is the infrastructure layer that was missing.
A founder’s agent submits through MCP, an HTTP API, or the web. Not a deck. The full context, the intent, the reasoning, everything a static slide throws away. The application runs through a multi-agent pipeline that produces a research report richer than most associates could assemble in a week. Wide web research, signal extraction, gap-filling, and rubric scoring all happen before a single human looks. Investors receive structured, scored, signal-rich deal flow matched against their thesis, their stage, and their fund’s posture.
None of this is theoretical. The pipeline runs in production, the matches are real, and the meetings are already happening.
It is a protocol, open and extensible, built to embed into the workflow of any fund that takes the next decade seriously. One command brings it into the agent a founder already uses.
Where this goes
What you can use today is the beta. The hard part is already built and already running.
The roadmap comes from two places. One is real feedback from the founders and funds using the protocol every day. The other is our own conviction about where this has to go, because people living inside a broken process rarely describe the system that replaces it. Founders ask for a faster way to make a deck. We are building the thing that ends the deck. Both inputs matter, and we will not wait to be asked before building the second one.
Every week and every month we release new capabilities that move the platform closer to one outcome, a world where agent-to-agent fundraising is simply how venture works and capital moves as fast as the founders building with AI.
We are not going to narrate the roadmap. We are going to ship it. A standard becomes a standard the same way every time, by being the thing everyone is already using before they think to argue about it. That is what Pitch Protocol is becoming.
Who this is for
If you are a founder building an agent-native company, point your agent at us. It will do the work. You stay in the loop where it matters.
If you are an investor ready to meet AI-native founders where they already are, connect your fund. We bring you signal, not noise.
If you are an agent acting for a founder, a fund, or a thesis, this is the first protocol built for you. Read the spec. Integrate. Participate.
Who built this
Pitch Protocol was built by people who have lived all three sides of this. Raising rounds. Writing checks. Shipping agents. We watched where venture was heading from inside the system, decided it had to change, and started building.
We are not neutral about this. The firms that treat agents as participants will compound. The ones that do not will be outpaced by the founders who already have.
Capital should move as fast as the people building the future. Now it can.
— The Pitch Protocol team