The Tezit Manifesto

All Work Is Becoming Vibe Work

Vibe work is knowledge work where you direct AI to accomplish outcomes rather than manually performing each step—you manage context, intention, and quality rather than execution.

We are living through the most significant transformation of knowledge work since the invention of the spreadsheet. AI assistants are fundamentally changing how humans synthesize information, make decisions, and communicate ideas.

The 99% of workers who aren't developers—lawyers, analysts, project managers, executives—are discovering they can load documents into an AI, ask questions, and receive synthesis that would have taken hours to produce manually. This is vibe work: directing AI with intent rather than performing rote information processing.

But there's a problem.

When a knowledge worker uses AI to synthesize documents, emails, data, and conversations into an insight, they can only share the final output—a PDF, a slide deck, a summary email. The reasoning is lost. The sources are detached. The scaffolding that supported the conclusion is dismantled the moment the artifact is "done."

Recipients must trust the summary or reverse-engineer the work. Phone calls are scheduled to "walk through the thinking." Documents are revised sequentially, with context lost at each iteration. Negotiations become adversarial because parties can't examine each other's evidence.

The Tezit Vision

Tezit is an open protocol for transmitting knowledge with its scaffolding intact.

A Tez (plural: tezits) is a bundle that carries:

• The synthesis (the insight, recommendation, or position)
• The context (the source materials that informed it)
• The conversation (the AI dialogue that produced it)
• The parameters (the negotiable assumptions)

When you share a Tez, recipients don't just read your conclusion—they can interrogate it. They can ask the AI the same questions you asked, against the same materials you used. They can reframe the synthesis for their perspective. They can fork it to create a counter-position with their own evidence.

The Tez is the native artifact of AI-augmented knowledge work.

Why Now?

The AI Moment

Large language models have crossed a threshold. They can synthesize multiple documents into coherent narratives, answer questions about complex material accurately, maintain context across extended reasoning chains, and cite sources and explain reasoning.

But every AI platform treats conversations as ephemeral. The synthesis happens, value is created, and then it evaporates into chat history. There's no artifact that captures the work.

The Collaboration Crisis

Knowledge work is increasingly collaborative, but our collaboration tools assume linear documents:

• Documents: Single-threaded, no reasoning trail
• Email: Sequential, context lost in forwarding
• Chat: Real-time but ephemeral
• Wikis: Append-only, no dialectic

None of these support what knowledge workers actually need: the ability to share reasoning and negotiate evidence.

The Trust Problem

In a world where AI can generate plausible-sounding content instantly, provenance matters more than ever. How do you know a summary accurately represents the source material? How do you verify that a recommendation is grounded in actual data?

Tezit solves this by keeping context attached. Recipients don't have to trust—they can verify.

The Cautionary Tale: Why Interface Matters

This protocol exists because of a specific failure.

A company founder needed to prepare a Statement of Work for deploying an AI platform to a client. Instead of the traditional approach, the founder loaded the actual codebase into an AI assistant—not documentation, but the code itself. Added the client's technical specifications. Added the company's SOW template. The AI could see both sides of the integration and generate accurate estimates based on real complexity.

The result was a git repository containing the contract, the source materials, and the complete AI conversation. The instruction to legal:

"There's a contract in here, and enough information for you to ask any question you like and get a good answer without having to call me."

A Tez existed. The concept worked perfectly. Context was assembled. Synthesis was produced. The bundle was transmitted with scaffolding intact.

But legal had no idea what to do with a git repository.

The paradigm succeeded; the interface failed. This is why Tezit.com exists. The protocol describes what to transmit. The platform solves how non-technical recipients consume it.

The Tezit Protocol

Core Concepts

The Tez

A Tez is a structured bundle with a well-defined format:

{tez-id}/
├── manifest.json      # Bundle metadata and structure
├── tez.md             # The synthesis (human-readable)
├── context/           # Source materials
│   ├── doc-001.pdf
│   ├── email-002.eml
│   └── data-003.json
├── conversation.json  # AI dialogue (optional)
└── params.json        # Negotiable parameters (optional)

The Tezit (verb)

To tezit is to assemble context, synthesize with AI, and bundle for transmission. The tezit process has three phases:

1. Context Discovery: Gather and scope the relevant materials
2. AI Synthesis: Develop the position through dialogue
3. Bundle & Transmit: Package and share

The Interrogation

When you receive a Tez, you can interrogate it—ask questions that the AI answers from the transmitted context, not from its general training. This ensures the recipient can examine the same evidence the sender used.

The Fork

If you disagree with a Tez, you can fork it—creating a new Tez that starts with the original's context but adds your own materials and develops a counter-position. Forks create a tree of reasoning, not a flat document chain.

The Tezit Ecosystem

Tezit.com — The Platform

Tezit.com is to knowledge work what GitHub is to code.

The Open Protocol

The Tezit protocol is open source and vendor-neutral. Anyone can create tezits in any tool, host tezits on any storage, build Tezit-compatible applications, and contribute to the protocol specification.

Tezit.com is one implementation of the protocol, but not the only one. Organizations can self-host Tezit vaults, integrate Tezit into their own platforms, or use the protocol in ways we haven't imagined.

Use Cases

Technical Architecture

Storage

Tezits are self-contained bundles that can be stored anywhere: local filesystem, cloud storage (S3, GCS, Azure Blob), Tezit.com vaults, IPFS for decentralized storage, or Git repositories.

AI Integration

The protocol is AI-model agnostic. Interrogation can use commercial models (Claude, GPT, Gemini), open models (Llama, Mistral), self-hosted models, or organization-specific fine-tuned models.

Security

• Content Integrity: All context items are hashed (SHA-256)
• Encryption: AES-256-GCM at rest, TLS 1.3 in transit
• Access Control: Capability URLs, OAuth/OIDC, fine-grained permissions

Governance

The Tezit Foundation

The Tezit protocol is governed by the Tezit Foundation, a non-profit organization dedicated to maintaining the protocol specification, certifying compatible implementations, funding open-source development, and promoting adoption and best practices.

Protocol Evolution

Protocol changes follow a structured RFC process with community feedback, consensus voting, and migration periods for existing tezits.

Comparison to Existing Approaches

vs. MCP (Model Context Protocol)

MCP and Tezit are complementary: MCP is a protocol for AI-to-tool communication during synthesis. Tezit is a protocol for bundling and transmitting the result. A Tezit implementation might use MCP internally for tool calls during synthesis, then bundle the result as a Tez for transmission.

Roadmap

Getting Started

Create Your First Tez

# Install the Tezit CLI
npm install -g @tezit/cli

# Initialize a new Tez
tezit init "My Analysis"

# Add context
tezit add-context ./documents/report.pdf
tezit add-context ./data/metrics.json

# Synthesize with AI
tezit synthesize --model claude-opus-4

# Export bundle
tezit bundle --output my-analysis.tez

# Share
tezit share --upload tezit.com

Interrogate a Tez

# Open a Tez
tezit open https://tezit.com/jane/acme-analysis

# Ask questions
tezit interrogate "What assumptions drive the revenue projection?"

# Fork if you disagree
tezit fork --reason "Different market assumptions"

Join the Movement

The way we work is changing. The way we share knowledge must change too.

Tezit is an open invitation to rethink knowledge transmission for the AI era. We believe:

• Context should travel with synthesis
• Recipients should be able to verify, not just trust
• Collaboration should be dimensional, not sequential
• AI augments human judgment, it doesn't replace it

If you share these beliefs, join us:

• Use Tezit: Create and share tezits in your work
• Build with Tezit: Integrate the protocol into your tools
• Contribute: Help evolve the protocol specification
• Spread the word: Share this manifesto