---
title: Ivy Biosciences Docs
description: >-
  Ask a biology question or call a tool directly. .bloom* runs the tools that
  answer it and returns results you can cite, export, and check.
tags:
  - overview
  - getting-started
  - platform
---

.bloom* is the Ivy Biosciences drug discovery platform. Ask a question in plain language, and it runs the tools that answer it: **protein structure**, **docking**, **ADMET**, **omics**, **synthesis**, and scientific databases.

Every run returns structured results and a record you can check. Re-query it, export it, or feed it into the next step.

> **Four ways in**

> **curl** and **chat** need no account for the public guest path. **MCP** and the **CLI** add saved sessions, higher quota, and the tools your plan allows once you sign in. All four reach the same .bloom* registry, so choose the one that fits how you work.

## Start here

Pick the closest; the sidebar carries the rest.

## Quickstart

    **Start without an account.** Try [guest chat](https://www.ivybiosciences.com/chat/new?guest=true), or paste the public [curl request](/docs/api#guest-curl) for an allowlisted run.

    **Pick your entry point.** Use chat for a question, curl for a quick request, CLI for batch work, or MCP when an AI coding agent should call the tools.

    **Run your first job.** Try the [ESMFold quickstart](/docs/esmfold) to fold a sequence, or the [DiffDock quickstart](/docs/diffdock) to dock a ligand.

    **Keep the result.** Sign in when you need saved sessions, higher quota, project history, or the tools your plan allows; then chain tools through [workflows](/docs/workflows) or the API.

## Install the CLI

**npm**

```bash
npm install -g @ivybiosciences/ivybloom-cli
ivybloom auth login --browser
```

**pnpm**

```bash
pnpm add -g @ivybiosciences/ivybloom-cli
ivybloom auth login --browser
```

**bun**

```bash
bun add -g @ivybiosciences/ivybloom-cli
ivybloom auth login --browser
```

Full install reference: [CLI install guide](/docs/cli).

## Explore by domain

Know the outcome but not the tool? [What you can achieve](/docs/capabilities) maps each goal to the tools that serve it.

## Workflows you can chain

## Reach it with curl or the API

The public orchestrator accepts a natural-language prompt over `curl` with no account. Authenticated API access adds persistent sessions, higher quota, and the full programmatic surface. The [API overview](/docs/api) explains both paths.

## Access and plans

> **Start free, then add persistence**

> The guest path starts with up to **$1 in compute per day** and does not need an account or card. Sign in for saved sessions and your own quota; see [Plans and tiers](/docs/plans-and-tiers) for current credits, model families, tool catalog, and API/MCP access.

For current rates, see [pricing](/pricing).

## Common questions

### Do I need an account to read the docs?

    No — the entire `/docs` tree is publicly indexed. Chat and the public curl orchestrator also work without an account within guest limits.

### What's the difference between Chat, curl, CLI, and MCP?

    Chat and curl are the fastest public entry points. Use the **CLI** for batch processing and reproducible pipelines, or **MCP** when a coding agent should call the registry. Sign in when you need saved sessions, higher quota, and the tools your plan allows. See [Get started](/docs/getting-started).

### Can I self-host?

    Through an enterprise arrangement today; ask [support](/help/support). The platform is licensed FSL-1.1-Apache-2.0, which converts to Apache 2.0 after 2 years, and a public source release is planned (the repositories are not public yet). Enterprise data providers and SLAs are arranged separately.

### Which tools support batch input?

    The guest path runs an allowlisted subset. Authenticated CLI and API workflows expose the catalog permitted by the account's plan; see [Plans and tiers](/docs/plans-and-tiers) and each tool's input page for limits.

### How are results stored and exported?

    Every job emits a structured result blob (JSON + tool-specific artifacts: PDB, SDF, CSV, PNG). Results render as canvas cards in Ivy and as JSON via `ivybloom jobs results <job_id>`. How long results are kept depends on your plan; see [Plans and tiers](/docs/plans-and-tiers).

### Where do I report a bug or request a tool?

    Open a ticket from [support](/help/support), or suggest a tool on the [suggestions board](/suggestions). Recent changes are listed in the [changelog](/changelog).

## Read next