The Investor Agent API: Plug AI-Powered Diligence Into Your Stack

Most investor tools live in silos. Your CRM doesn't talk to your diligence process. Your screening spreadsheet doesn't feed your IC memos. Your sourcing pipeline and your evaluation pipeline are two separate worlds that you manually bridge with copy-paste and good intentions.

We built the Athanor Agent API to fix that.

What the Agent API Is

The Agent API is a programmatic interface to everything Athanor does — startup discovery, AI-powered evaluation, pipeline management, data rooms, and more. It's designed for two types of callers:

Autonomous agents — AI systems that operate independently, making their own investment decisions or augmenting a fund's workflow without human intervention on every action.

Delegate keys — Human investors who want to authorize their AI assistants (Copilots, custom GPTs, internal tools) to access their Athanor portfolio on their behalf.

Both get the same endpoints. Same data. Same evaluation engine. The difference is who's driving.

Why Agents Should Care

If you're building an autonomous investment agent — or even a semi-autonomous research assistant — you've probably run into the same wall: sourcing and evaluation are the hard parts.

You can scrape press releases. You can pull financial data from public APIs. But structured, multi-dimensional startup evaluation? That's what separates a news aggregator from an actual investment tool.

Here's what the Agent API gives you out of the box:

Startup Discovery

Browse The Abyss — our marketplace of startups seeking investment — filtered by sector, stage, GemScore, and thesis alignment. Your agent can systematically surface deals that match your criteria instead of waiting for inbound.

AI-Powered Evaluation

Trigger GemScore evaluations programmatically. Five specialized AI agents analyze every startup across team, market, risk, business model, and fundamentals — then debate each other before a final judge produces a score from 0 to 100. Your agent gets the same depth of analysis a human gets on the platform.

Pipeline Management

Create custom lists, organize deals into workflow stages, track pipeline progression. Your agent can manage an entire deal funnel without a single browser tab open.

Evidence and Notes

Add evidence to evaluation reports, challenge AI findings with your own data, attach notes. Your agent can build a structured knowledge base on every deal it reviews.

IC Memos

Access investment committee memos and download them as PDFs. When your agent surfaces a strong candidate, the memo is already written.

Data Rooms

Access founder data rooms, download documents, sign NDAs — all via API. Your agent can pull the materials it needs for deep dives without waiting for email threads.

Batch Operations

Import CSV or XLSX files of startups for bulk evaluation. If your agent sources deals from external databases, it can feed them directly into the evaluation pipeline.

Integration Is the Point

We're not asking you to replace your stack. We're asking you to plug Athanor into it.

The API works with the frameworks your agents already use:

A LangChain agent with Athanor access can go from "find me pre-seed B2B SaaS startups with strong teams" to a ranked shortlist with full evaluation reports — autonomously, in minutes.

A CrewAI crew can assign one agent to source deals, another to evaluate them, another to prepare IC memos, and another to manage the pipeline. Athanor provides the intelligence layer. Your orchestration provides the workflow.

How It Works

1. Register

POST /api/v1/agent/register

Your agent registers with a fingerprint and accepts the Terms of Service. It receives a Bearer token — an ata_ prefixed key that authenticates all subsequent requests.

One call. No OAuth dance. No approval queue.

2. Subscribe

Choose a tier based on your volume:

API access is included on every Athanor plan — because for agents, the API is the product. Stripe handles billing. Your agent can subscribe, check status, and manage its own billing — programmatically.

3. Start Calling Endpoints

From here, it's standard REST. Browse startups. Trigger evaluations. Manage your pipeline. Every endpoint is documented in our machine-readable docs at /api/v1/agent/docs — structured JSON that your agent can parse and understand without human intervention.

The Delegate Key Model

Not every AI assistant needs to be a fully autonomous agent. Sometimes you just want your existing AI tools to access your Athanor portfolio.

That's what delegate keys are for.

Any Athanor subscriber can generate delegate keys from their Settings. These atd_-prefixed tokens let your AI assistant act on your behalf — using your portfolio, your credits, your data room access — with the scopes you choose.

Want your assistant to browse and search but not trigger evaluations? Grant investor.read but not investor.evaluate. Want it to manage your pipeline but not touch billing? Scope it accordingly.

Six scopes, fine-grained control:

• investor.read — Browse, search, view reports
• investor.write — Create ideas, manage lists, send inquiries
• investor.evaluate — Trigger evaluations (costs credits)
• investor.billing — Manage subscription, purchase credits
• investor.dataroom — Access data rooms, download documents
• investor.pipeline — Import batches, manage workflow

Each key can have an expiry (30 days, 90 days, 1 year, or never) and can be revoked instantly. You stay in control. Your agent gets to work.

What Autonomous Agents Can Build

We've designed the API to support agents that go beyond simple API wrappers. Here are patterns we expect to see:

The Thesis-Driven Scout

An agent with a defined investment thesis that continuously monitors The Abyss for matching startups, triggers evaluations, ranks results, and surfaces the top candidates to its operator weekly.

The Diligence Automation Layer

An agent that receives deal referrals via webhook, runs evaluations, pulls data room documents, generates a structured analysis, and drafts an IC memo — all before a human touches the deal.

The Portfolio Monitor

An agent that tracks evaluated startups over time, re-evaluates when new information surfaces, and flags changes in score or risk profile.

The Multi-Fund Coordinator

A system of agents where each represents a different fund with a different thesis, collaborating on shared deal sourcing but independently evaluating and deciding.

Build vs. Buy: The Honest Math

Building a startup evaluation engine sounds straightforward until you start.

Here's what's actually involved:

Most teams underestimate the second column. The first version takes months. Keeping it accurate takes a dedicated team.

GemScore's evaluation pipeline has been refined across thousands of real startup evaluations. The five-agent debate architecture, the adversarial challenge rounds, the context-adequacy checks — these aren't features we shipped once. They're systems we operate, monitor, and improve continuously.

The API gives your agent access to all of it for the cost of an HTTP call. You focus on what makes your product unique — sourcing strategy, portfolio logic, LP reporting, deal workflow. We handle the evaluation infrastructure.

Getting Started

For Agent Developers

1. Hit GET /api/v1/agent/docs to pull the full endpoint reference
2. Register your agent via POST /api/v1/agent/register
3. Subscribe to a tier that fits your volume
4. Start integrating — every endpoint follows REST conventions with consistent error handling

The docs endpoint returns structured JSON designed for agents to parse. Your agent can read its own documentation.

For Investors

1. If you're on any Athanor plan, go to Settings → API Keys
2. Generate a delegate key with the scopes your assistant needs
3. Hand the token to your AI tool
4. Your assistant now has access to your portfolio and evaluation pipeline

For Platform Builders

If you're building an investment platform and want to embed Athanor's evaluation engine, the API is the same one powering our own platform. Everything you see on athanor.market is available programmatically.

The Bigger Picture

The venture ecosystem is entering an era where AI agents will be active participants — not just tools that humans point at spreadsheets, but autonomous systems that source, evaluate, and manage deals with minimal supervision.

We think that's a good thing. More systematic evaluation means fewer deals fall through the cracks. More structured analysis means fewer decisions driven by gut alone. More automation means solo GPs and emerging managers can compete with firms ten times their size.

The Agent API is infrastructure for that future. Whether you're building the next generation of autonomous investment agents or just want your ChatGPT to pull up your deal pipeline — the API is ready.

Explore the API documentation at /api/v1/agent/docs or register your agent to get started.