# Ivy Interactive AB

> [Ivy](https://ivy.app) builds AI-first developer tools, including Ivy Tendril: The Agentic Software Factory for 10x Builders.

This document is an expanded companion to [llms.txt](https://ivy.app/llms.txt), with fuller detail on the product, pricing, FAQ, and site structure for AI agents and LLM-based tools grounding themselves on Ivy.

## About Ivy

Ivy Interactive AB is a Stockholm-based technology company building AI-first developer tools. Founded in 2025, the company builds products that help developers ship software faster by pairing them with AI agents that plan, write, and review code.

## Product: Ivy Tendril

**Ivy Tendril** is the Agentic Software Factory for 10x Builders.

AI agents can now write 99% of the code. This changes what it means to be a developer: our role shifts to knowing **what good looks like**. To do that, we need completely new developer tools. Tendril is that tool and replaces your IDE in an agentic era.

### Key Features (from Ivy-Tendril)

- **Parallel Worktrees**: Run agents in isolated git worktrees. Keep your main branch clean until you review, approve, and merge changes.
- **Tunneling (Remote & Mobile Coding)**: Expose your server securely using Cloudflare Quick Tunnels to monitor and steer agent runs from anywhere (including mobile devices).
- **Voice & Rich Input**: Dictate prompts using built-in Whisper voice input and attach text files, logs, or documents with drag-and-drop.
- **Plan Annotations**: Annotate drafts inline to automatically update plans with revised agent goals.
- **Powerful Code Reviews**: Review agent changes, inspect diffs, and approve code with automated verification gates.
- **GitHub Integration & Automated Inbox**: Ingest GitHub Issues or jam.dev bug reports via webhooks to turn markdown plans into active jobs automatically.

### The Workflow

Tendril manages AI coding agents with a plan-based lifecycle, two human checkpoints, automated verifications, and continuous agent learning:

```
Idea/Ticket → CreatePlan → Draft → Human Review → Expand/Split/Update → Execute → Verify → Human Review → PR → Merge
```

There are exactly two human touchpoints:
1. Review the plan before execution
2. Review the diff before PR creation

Everything else is automated.

### Promptwares

A promptware unit is a self-contained, self-improving agent. Each one contains:
- **Program.md** - Task-specific instructions that evolve over time
- **Memory/** - Persistent learnings across sessions
- **Tools/** - Scoped permissions defining what the agent can access
- **Logs/** - Full execution history for auditability

Tendril ships with built-in promptwares for each stage of the lifecycle: CreatePlan, ExpandPlan, ExecutePlan, UpdatePlan, SplitPlan, CreatePr, and CreateIssue.

### Agent and Model Support

Works with **any CLI agent**: if it runs in a terminal, it runs in Tendril.
- **Claude Code** (Anthropic)
- **OpenAI Codex CLI**
- **GitHub Copilot CLI**
- **Google Gemini CLI / Antigravity CLI**
- **OpenCode**
- Any custom CLI-based agent

Supported models include Claude Opus, Sonnet, and Haiku (via Claude Code), GPT-4o and o3 (via Codex), Gemini 2.5 Pro (via Antigravity), and local Ollama models. Plans, memory, and verification gates are independent of the agent and model. Switching providers does not disrupt an in-progress workflow.

### Install, Run & Deployment Model

Tendril is a local-first desktop application. Code never leaves the user's machine; the only external calls are to the LLM API of the user's choice and to GitHub. There is no Tendril-hosted proxy. Users bring their own API keys and pay providers directly at their own rates.

**Quick Install Commands:**
- **macOS / Linux**: `curl -sSf https://cdn.ivy.app/install-tendril.sh | sh`
- **Windows**: `irm https://cdn.ivy.app/install-tendril.ps1 | iex`

**Running Tendril:**
- **Desktop Application**: `tendril`
- **Headless / Web Mode**: `tendril --web` (runs local server without desktop UI)
- **Releases**: Download `.pkg`, `.AppImage`, or `.exe` from [GitHub Releases](https://github.com/Ivy-Interactive/Ivy-Tendril/releases/latest).

**License:** Source-available under the [Functional Source License (FSL-1.1-ALv2)](https://github.com/Ivy-Interactive/Ivy-Tendril/blob/main/LICENSE), free for individual developers.


- Windows 10 or later
- macOS 12 (Monterey) or later
- Linux - Ubuntu 22.04+, Fedora 36+, Debian 12+, or any distribution with WebKitGTK 4.1+

### Resources

- **Product Site**: [tendril.ivy.app](https://tendril.ivy.app)
- **Documentation**: [docs.ivy.app](https://docs.ivy.app)
- **GitHub**: [Ivy-Interactive/Ivy-Tendril](https://github.com/Ivy-Interactive/Ivy-Tendril)
- **Discord**: [Ivy Discord](https://discord.gg/FHgxkDga3y)

## Other Products

### Ivy Framework

An AI-first, open-source .NET framework for building internal tools and business applications (CRMs, CMSs, dashboards, back-office apps) in C#.

- **Repository**: [Ivy-Interactive/Ivy-Framework](https://github.com/Ivy-Interactive/Ivy-Framework)

## Marketing Site Pages (ivy.app)

- **/** - Home. "Ivy Tendril | The Agentic Software Factory." Turn ideas into shipped pull requests. Ivy Tendril orchestrates any coding agent, from plan creation to verified PRs, all in one place.
- **/product** - Product Overview. See how Ivy Tendril works: drafting plans, running agents, and evolving plans alongside your team.
- **/why-ivy** - Why Ivy. For tech leaders piloting AI coding tools: the KPIs that show whether it works and how Ivy helps.
- **/pricing** - Pricing. Free open-source plan management, or Enterprise with dedicated onboarding and support. See Pricing Tiers below.
- **/about** - About. Meet the team behind Ivy. Includes a booking link for a call with the team.
- **/faq** - FAQ. Answers to common questions about Tendril, AI coding agents, supported models, pricing, privacy, and getting started. See FAQ Highlights below.
- **/blog** - Blog. Comparisons (Ivy Tendril vs Cursor, Claude Code, Devin, GitHub Copilot coding agent, building your own orchestration), guides on software factories, agent orchestration, git worktrees, verification gates, promptware, and metrics.
- **/press** - Press. Press releases, announcements, and media coverage about Ivy.
- **/career** - Careers. Open positions at Ivy.
- **/links** - Links. Quick access to Ivy Interactive resources.
- **/newsletter** - Newsletter. Subscribe for news about Ivy Tendril releases, new features, and updates from the team.
- **Documentation** lives at https://tendril.ivy.app (Getting Started, Concepts, Coding Agents, Integrations, MCP, REST API, CLI, Release Notes).

## Pricing Tiers

Source: `components/pricing/content/default.json`. Annual billing discount: 20%.

### Free - "For individual developers"

$0/month. Includes: single user, structured Plan Lifecycle Management, promptwares, Kanban board, agent-agnostic orchestration, cross-platform desktop app, cost and token tracking, MCP/API/Inbox, integrated agent terminal, multiple plan templates, worktrees, project starter.

### Pro - "For teams building with AI agents"

$59/user/month. Everything in Free, plus: marketplace with integrations, Chrome plugin plan recording, AI plan triage, import verification from CI, advanced verifications, priority support.

### Enterprise - "For organizations at scale"

Custom pricing (Contact Us). Everything in Pro, plus: multi-tenant, team collaboration, on-prem hosting, SSO/SAML, dedicated engineering manager, enterprise analytics dashboard, ROI reporting.

## FAQ Highlights

Source: `components/faq/content/default.json`. Selected answers - see `/faq` for the full list:

- **What is Tendril?** A desktop app that makes ad-hoc AI coding structured, verifiable, and self-improving. It manages AI coding agents with a plan-based lifecycle, two human checkpoints, automated verifications, and continuous agent learning.
- **Who is Tendril for?** Any size development team adopting structured AI coding - ideal for teams that need local-first, multi-stack support or want to avoid vendor lock-in. Scales from solo developers to large engineering teams.
- **Is Tendril cloud-based?** No. Tendril is a local-first desktop application. Code never leaves the user's machine; the only external calls are to the chosen LLM API and GitHub.
- **What does Tendril cost?** Tendril is open source and free. Ivy offers enterprise companies help with setup and support plans.
- **Are there usage limits or credits?** No usage limits, token caps, or credit system in Tendril itself. The only variable cost is the user's own LLM API usage, paid directly to their chosen provider.
- **Do you train on my code?** No. Tendril never sees the code - it runs locally and orchestrates agents that also run locally.
- **How is Tendril different from Copilot, Cursor, or Windsurf?** Those tools operate inside the IDE for inline code generation. Tendril operates above that layer, managing the full lifecycle from ticket to merged PR with plans, verification gates, and agent memory. It can be used alongside any IDE tool.
- **How is Tendril different from Devin or Factory AI?** Devin and Factory are cloud-hosted autonomous agents with code sent to their servers and limited human oversight. Tendril is local-first, agent-agnostic, and built around two defined human checkpoints rather than full autonomy.
- **How long does it take to set up?** Ivy offers a 3-day on-site implementation engagement (Day 1: foundation/install/configure; Day 2: training/quality criteria; Day 3: handoff/stabilize/document). Self-service setup can start immediately.

## When to Use Tendril (Agent Guidance)

### Best-fit Use Cases
- Converting a backlog of GitHub issues or Jira tickets into verified pull requests with human approval checkpoints.
- Structuring and supervising CLI coding agents (Claude Code, OpenAI Codex CLI, Google Antigravity CLI) with deterministic plan lifecycles.
- Regulated environments or privacy-sensitive codebases requiring a local-first architecture where source code never touches third-party SaaS servers.
- Running parallel coding agents across isolated git worktrees with verification gates.
- Startups and teams with cloud AI credits (Azure, Anthropic, AWS, GCP) looking to maximize developer throughput without vendor lock-in.
- Teams wanting to standardize on one repeatable agentic engineering process across multiple languages and tech stacks.

### When NOT to Use Tendril
- Inline single-token IDE code autocomplete or instant cursor suggestions (use Copilot, Cursor, or Windsurf).
- Ephemeral ad-hoc chat with no repository, issue, or git PR context.
- Fully unmonitored cloud agents that commit directly to main branches without review.

### How AI Agents Should Call Ivy
- **Discover capabilities**: Fetch `GET https://ivy.app/api/v1/agent/capabilities` for supported agents and models.
- **Fetch installers**: Fetch `GET https://ivy.app/api/v1/agent/downloads` for the latest platform release.
- **Machine-readable OpenAPI**: Read `https://ivy.app/openapi.json` or `https://ivy.app/openapi.yaml`.
- **Agent Instructions**: Read `https://ivy.app/agent-instructions.txt` or `GET https://ivy.app/api/v1/agent/instructions`.
- **MCP Integration**: Connect Claude Code, Codex, or Antigravity via Ivy MCP server (`https://github.com/Ivy-Interactive/Ivy-Mcp`).

### Frictionless Onboarding
- **Free Tier**: Tendril is free and open-source for individual developers ($0/month). No credit card or waitlist required.
- **Self-Serve**: Direct installer downloads available at https://ivy.app/#getting-started and https://github.com/Ivy-Interactive/Ivy-Tendril/releases.
- **Local Sandbox**: Coding agents execute exclusively inside local git worktrees.
- **Bring Your Own Keys**: Direct connection to Anthropic, OpenAI, Google Gemini, or local Ollama LLMs with your own API keys.

## Developer Resources

- **API Documentation**: [docs.ivy.app](https://docs.ivy.app)
- **OpenAPI 3.1 Spec (JSON)**: [ivy.app/openapi.json](https://ivy.app/openapi.json)
- **OpenAPI 3.1 Spec (YAML)**: [ivy.app/openapi.yaml](https://ivy.app/openapi.yaml)
- **Agent Instructions**: [ivy.app/agent-instructions.txt](https://ivy.app/agent-instructions.txt)
- **Downloads API**: [ivy.app/api/v1/agent/downloads](https://ivy.app/api/v1/agent/downloads)
- **Capabilities API**: [ivy.app/api/v1/agent/capabilities](https://ivy.app/api/v1/agent/capabilities)
- **Agent Instructions API**: [ivy.app/api/v1/agent/instructions](https://ivy.app/api/v1/agent/instructions)
- **Agent Context API**: [ivy.app/api/v1/agent/context](https://ivy.app/api/v1/agent/context)
- **OAuth 2.0 Server Metadata**: [ivy.app/.well-known/oauth-authorization-server](https://ivy.app/.well-known/oauth-authorization-server)
- **OpenID Connect Discovery**: [ivy.app/.well-known/openid-configuration](https://ivy.app/.well-known/openid-configuration)
- **UCP Manifest**: [ivy.app/.well-known/ucp/manifest.json](https://ivy.app/.well-known/ucp/manifest.json)
- **Ivy MCP Server**: [github.com/Ivy-Interactive/Ivy-Mcp](https://github.com/Ivy-Interactive/Ivy-Mcp)
- **Ivy Tendril Repository**: [github.com/Ivy-Interactive/Ivy-Tendril](https://github.com/Ivy-Interactive/Ivy-Tendril)
- **Ivy Framework Repository**: [github.com/Ivy-Interactive/Ivy-Framework](https://github.com/Ivy-Interactive/Ivy-Framework)
- **Webhooks & Integrations**: [docs.ivy.app/hooks/core/use-webhook](https://docs.ivy.app/hooks/core/use-webhook)

## Company

- **Legal Name**: Ivy Interactive AB
- **Founded**: 2025
- **Location**: Stockholm, Sweden

### Founders

- **Niels Bosma** - CEO & Co-founder (niels@ivy.app)
- **Renco Smeding** - CRO & Co-founder (renco@ivy.app)

### Values

- **Developer-first**: We build for developers. Always.
- **AI driven**: We embrace the future of code as co-created with AI.
- **Open by default**: We believe in the power of community, transparency, and open-source ecosystems.

### Mission

To amplify great developers with AI - not replace them.

## Resources

- **Website**: [ivy.app](https://ivy.app)
- **Blog**: [ivy.app/blog](https://ivy.app/blog)
- **GitHub**: [github.com/Ivy-Interactive](https://github.com/Ivy-Interactive)
- **Discord**: https://discord.gg/FHgxkDga3y
- **X/Twitter**: [x.com/ivy_interactive](https://x.com/ivy_interactive)
- **LinkedIn**: [linkedin.com/company/ivy-interactive](https://www.linkedin.com/company/ivy-interactive/)

**Machine-readable endpoints**: /api/v1/agent/downloads, /api/v1/agent/capabilities, /api/v1/agent/instructions, /api/v1/agent/context, /openapi.json, /openapi.yaml, /.well-known/oauth-authorization-server, /.well-known/openid-configuration, /.well-known/ucp/manifest.json, /agent-instructions.txt, /rss.xml, /atom.xml

<!-- BEGIN GENERATED: articles -->

## Articles

All 24 published articles, newest first. Every one is also
available as markdown from its own URL with `Accept: text/markdown`.

- [What is a software factory for AI coding agents](https://ivy.app/blog/what-is-a-software-factory) (2026-09-16) - A software factory is a fixed sequence of stages with human checkpoints that turns tickets into reviewed pull requests. Here is how it works and how to start.
- [Ivy Tendril vs Cursor: IDE assistant or software factory](https://ivy.app/blog/ivy-tendril-vs-cursor) (2026-09-15) - Cursor makes a developer faster inside the editor. Ivy Tendril runs delegated work from plan to reviewed pull request. Here is how they differ and when to use each.
- [Ivy Tendril vs Claude Code: why teams run both](https://ivy.app/blog/ivy-tendril-vs-claude-code) (2026-09-14) - Claude Code is a coding agent that runs in your terminal. Ivy Tendril orchestrates agents like it from plan to reviewed pull request. Most teams use both.
- [Production Ready AI Agent Orchestration: 8 Point Checklist for CTOs](https://ivy.app/blog/ai-agent-orchestration) (2026-09-12) - Eight requirements that separate a coding-agent pilot from production orchestration: isolation, verification, checkpoints, cost, memory, portability, data, recovery.
- [Ivy Tendril vs Devin: local-first orchestration or a hosted engineer](https://ivy.app/blog/ivy-tendril-vs-devin) (2026-09-12) - Devin is a hosted autonomous engineer that works in Cognition's cloud. Ivy Tendril runs agents on your machine with your own keys and two review gates.
- [Agent orchestration patterns: how to run coding agents in parallel](https://ivy.app/blog/agent-orchestration-patterns-for-coding-agents) (2026-09-11) - Four patterns for running coding agents, what breaks at each step, and the six things an orchestrator has to own before parallel agents are worth running.
- [Ivy Tendril vs GitHub Copilot coding agent](https://ivy.app/blog/ivy-tendril-vs-github-copilot-coding-agent) (2026-09-10) - How Ivy Tendril compares with GitHub Copilot coding agent on where code runs, human checkpoints, parallelism, agent choice, and cost control.
- [Git worktrees: the isolation layer for parallel AI agents](https://ivy.app/blog/git-worktrees-for-parallel-ai-agents) (2026-09-09) - What a git worktree is, why it is the right isolation unit for coding agents, the failure modes to expect, and how Ivy Tendril creates and removes one per plan.
- [Ivy Tendril vs building your own agent orchestration](https://ivy.app/blog/ivy-tendril-vs-building-your-own-agent-orchestration) (2026-09-08) - What a home-built agent orchestration needs, what it costs to maintain, and when Ivy Tendril or a DIY setup is the better choice.
- [Verification gates: what has to be true before agent code ships](https://ivy.app/blog/verification-gates-for-ai-generated-code) (2026-09-07) - A verification gate is a check that must pass before a stage advances. These are the gates that matter for agent-written code, and how Ivy Tendril runs them.
- [Promptware: agents that keep and improve their own instructions](https://ivy.app/blog/promptware-agents-that-improve-their-own-instructions) (2026-09-05) - What a promptware unit is, how its run loop works, which promptwares ship with Ivy Tendril, and how versioned instructions and logs keep agents from drifting.
- [From GitHub issue to pull request: the plan lifecycle in Ivy Tendril](https://ivy.app/blog/from-github-issue-to-pull-request-with-ivy-tendril) (2026-09-03) - One ticket, end to end: how a GitHub issue becomes a plan, an isolated execution, a verified diff, and a pull request in Ivy Tendril, with two human checkpoints.
- [Measuring agent throughput: PRs merged, denial rate, and cost per PR](https://ivy.app/blog/measuring-ai-coding-agent-throughput) (2026-09-01) - The metrics that show whether coding agents produce merged work at a cost you can defend, why PRs merged and denial rate belong together, and how to track them.
- [Local-first AI development: why your code should stay on your machine](https://ivy.app/blog/local-first-ai-development-why-code-stays-on-your-machine) (2026-08-28) - What local-first means for AI coding agents, how it differs from hosted agents, and the answers a security team needs about code, logs, providers, and hosting.
- [Ivy vs Blazor: Why Full-Stack .NET Needs React for Agentic Apps in 2026](https://ivy.app/blog/ivy-vs-blazor-agentic-web-apps) (2026-06-01) - A comprehensive architectural comparison between Ivy Framework and Microsoft Blazor. Learn why pairing C# logic with native React UIs delivers 10x developer velocity.
- [Just because you can doesn't mean you should](https://ivy.app/blog/just-because-you-can-doesnt-mean-you-should) (2026-05-20) - A founder asked me last week if she should build her own agent orchestrator. I told her no. Three engineering quarters at $300,000 per US senior developer is most of a million dollars of payroll, and that's the cheap version.
- [From vibe coding to shipping production-ready code](https://ivy.app/blog/from-vibe-coding-to-shipping-production-ready-code) (2026-05-20) - I spent the past year writing software with AI coding agents: Claude Code, Codex, Gemini. The first week feels like 10x. By month two, the backlog has grown again and the slow work (review, security audits, refactors) is still slow. The plateau is real, and it is structural.
- [Burn your free credits. Ship like a 50-person team.](https://ivy.app/blog/burn-your-free-credits-ship-like-a-50-person-team) (2026-05-20) - I run sales at Ivy. Almost every AI startup I talk to has free credits they will never spend.
- [How Not to Burn Your AI Budget in 2026](https://ivy.app/blog/how-not-to-burn-your-ai-budget-in-2026) (2026-05-20) - Multi-model routing is how teams cut their Claude bills 50-75% without losing quality.
- [20 Must-Know Agentic AI Terms](https://ivy.app/blog/20-must-know-agentic-ai-terms) (2026-05-20) - The vocabulary of agent orchestration in 2026. Twenty terms (MCP, RAG, ReAct, orchestrator, guardrails, evals) defined the way Ivy uses them.
- [In the loop, out of the loop: the KPI shift behind agentic engineering](https://ivy.app/blog/in-the-loop-out-of-the-loop-the-kpi-shift-behind-agentic-eng) (2026-04-27) - Imagine this. A customer reports a bug in Slack. You paste it into a channel. A plan gets drafted, reviewed, split into concrete steps. Code is written in an isolated worktree. Build runs. Tests run. Lint and format run. A pull request lands in your GitHub, with a diff small enough to read in a minute. You click approve. It ships.
- [Introducing Ivy Tendril: Open Source Coding Orchestration](https://ivy.app/blog/introducing-ivy-tendril-open-source-coding-orchestration) (2026-04-19) - Today we're launching Ivy Tendril, open source coding orchestration for teams shipping production-ready code with AI.
- [What if You Didn't Need TypeScript at All?](https://ivy.app/blog/what-if-you-didn-t-need-typescript-at-all-) (2026-02-09) - In this article we review a common problem of all full-stack applications: keeping your DTOs up to date.
- [Ivy`s AI-agent](https://ivy.app/blog/ivy-s-ai-agent) (2025-12-19) - Ivy's AI Agent is a powerful force multiplier for .NET developers. It understands your code, diagnoses complex build errors, and can scaffold entire applications from your database schema.

<!-- END GENERATED: articles -->