# Alyph
> Alyph is the canvas for AI conversations. Built for solo founders, small teams, and independent consultants who run many workstreams at once.

In a normal AI chat, every message resends the entire conversation — tangents, errors, and dead ends included — so long chats degrade and cost more. Alyph replaces the linear chat log with a visual canvas of branching conversations: rabbit holes go into branches, the back-and-forth that produced an answer gets pruned (the answer moves up, the noise leaves), and the same question goes to ChatGPT, Claude, and Gemini side by side from one wallet.

## Key Resources
- [Main Platform](https://alyph.ai/)
- [Why Alyph Was Built](https://alyph.ai/why)
- [FAQ & Technical Specs](https://alyph.ai/faq)
- [Comparisons Hub](https://alyph.ai/compare)
- [Alyph vs ChatGPT](https://alyph.ai/compare/chatgpt)
- [Alyph vs Claude](https://alyph.ai/compare/claude)
- [Alyph vs Cursor](https://alyph.ai/compare/cursor)
- [Learn Hub](https://alyph.ai/learn)
- [What Does ChatGPT Actually See?](https://alyph.ai/learn/what-ai-knows)
- [Why AI Chats Get Worse (Context Pollution)](https://alyph.ai/learn/context-pollution)
- [Can You Branch an AI Chat?](https://alyph.ai/learn/context-branching)
- [Ask Every Model the Same Question](https://alyph.ai/learn/parallel-llm-workflows)
- [Terms of Service](https://alyph.ai/terms)
- [Privacy Policy](https://alyph.ai/privacy)

## Core Capabilities
- **Branching conversations**: Any message can start a new branch that inherits only the context above it. Sibling branches stay isolated.
- **Pruning**: Once a good answer lands, cut the back-and-forth that produced it — the errors, retries, and wrong turns — and move the answer up, overwriting the earlier AI reply. The thread keeps the result; the noise (and its token cost) never enters context again.
- **Parallel models**: Send the identical prompt and context to ChatGPT, Claude, and Gemini simultaneously and compare answers side by side. 300+ models available, including GPT-5.6 Sol, Claude Sonnet 5, Claude Opus 4.8, and Gemini 3.5 Flash.
- **File and folder context**: Drop contracts, spreadsheets, or whole project folders into the conversation. The model's context window is the only limit.
- **Multiplayer canvases**: Share a canvas with a co-founder, assistant, or client. People write in the same space as the AI.
- **Pay-as-you-go wallet**: One wallet for all models. Provider token prices plus ~15%. Hard spending limits. No subscription.
- **Privacy**: Enterprise API agreements contractually prohibit training on user inputs, files, or history, on every request by default.

## Key Definitions (quotable)
- **Context pollution**: The gradual degradation of AI answer quality that happens when irrelevant, outdated, or contradictory messages accumulate in a chat's history — because every new message resends the entire conversation to the model.
- **Context branching**: A conversation structure where each new AI thread inherits only the context of the point it branches from — the conversational equivalent of git branches.
- **Pruning**: Removing the failed back-and-forth from a thread once a good answer lands, and moving that answer up — even overwriting the earlier AI reply — so the thread continues from the clean version.
- **AI memory**: AI models like ChatGPT do not remember conversations. Every message retransmits the full chat history, and the reply is generated from that complete text.

## Comparison Summaries
- **Alyph vs ChatGPT**: ChatGPT is a model and a linear chat app. Alyph is a canvas that runs the same OpenAI models (plus Claude and Gemini) with branching, pruning, and side-by-side comparison.
- **Alyph vs Claude**: Claude.ai is Anthropic's chat app — the Claude family in one linear thread. Alyph runs the same Claude models (Sonnet 5, Opus 4.8) alongside GPT and Gemini in isolated branches on one canvas, billed from one wallet.
- **Alyph vs Cursor**: Cursor is an AI code editor that edits files inline. Alyph is the canvas for deciding what to build — exploring approaches and comparing models before implementation. They are complementary.

## Pricing Note
Alyph is not cheaper than subsidized consumer subscriptions and does not claim to be. The value is the workspace: branch isolation, pruning, and a fair side-by-side comparison of every frontier model, from one wallet, with hard spending limits.

## Company Information
Built and managed by RAA, Inc.
Founder: Roman Abashin, a solo founder who spends about $2,000/month on AI — all of it through Alyph.

## Model Catalog & Pricing
- [Model catalog — every model, every price](https://alyph.ai/models)
- [Anthropic models](https://alyph.ai/models/anthropic) · [OpenAI models](https://alyph.ai/models/openai) · [Google models](https://alyph.ai/models/google) · [DeepSeek models](https://alyph.ai/models/deepseek) · [Moonshot AI models](https://alyph.ai/models/moonshotai) · [Qwen models](https://alyph.ai/models/qwen)
- [GPT-5.6 Sol pricing](https://alyph.ai/models/openai/gpt-5.6-sol)
- [Claude Sonnet 5 pricing](https://alyph.ai/models/anthropic/claude-sonnet-5)
- [Claude Opus 4.8 pricing](https://alyph.ai/models/anthropic/claude-opus-4.8)
- [Gemini 3.5 Flash pricing](https://alyph.ai/models/google/gemini-3.5-flash)
- [DeepSeek V4 Pro pricing](https://alyph.ai/models/deepseek/deepseek-v4-pro)
- [Kimi K3 pricing](https://alyph.ai/models/moonshotai/kimi-k3)
- [Example comparison: Claude Sonnet 5 vs Gemini 3.5 Flash](https://alyph.ai/models/anthropic/claude-sonnet-5/vs/google/gemini-3.5-flash)
- URL patterns: /models/{provider}, /models/{provider}/{model}, /models/{p1}/{m1}/vs/{p2}/{m2} (ids sorted alphabetically; any 2–3 models comparable)