# 07 — Brand Messaging Publication status: launch-ready draft. Token-related language requires board and legal review before publication. ## Core positioning ### Category Primary category: Verifiable AI Memory Infrastructure Plain-English category: The memory card for AI Technical category: Provider-neutral AI memory and proof layer Enterprise category: Auditable AI memory control plane ### Positioning statement For AI users, developers, and enterprises who need durable memory without provider lock-in, Enigma is provider-neutral AI memory/proof infrastructure. Enigma creates portable Memory Passports, keeps canonical memory local-first or customer-controlled, injects only scoped context into model and tool calls, and emits signed receipts for Enigma-mediated memory lifecycle and boundary events. ### One-liner Enigma is the memory card for AI: portable context, local-first custody, and offline-verifiable receipts across models and agents. ### One-paragraph boilerplate Enigma is provider-neutral infrastructure for AI memory and proof. It gives users and organizations portable Memory Passports, local-first or customer-controlled vaults, scoped context injection into model calls, and signed receipts for Enigma-mediated memory lifecycle and boundary events. Enigma helps AI remember without locking memory inside one provider, and helps people verify what happened without exposing raw memory or trusting a vendor dashboard. ### SanDisk analogy guidance Use the SanDisk comparison to explain category, not to imply hardware, consumer electronics, or storage-only functionality. The approved meaning is: SanDisk made storage portable across devices; Enigma makes AI memory portable across supported AI clients, agent workflows, and provider boundaries. Follow the analogy immediately with the proof layer: unlike a generic storage card, Enigma also emits receipts for declared Enigma-mediated memory events. Approved analogy sentence: “SanDisk made storage portable across cameras and computers; Enigma makes AI memory portable across models and agents, with receipts for what the memory layer did.” Avoid analogy drift: do not imply Enigma manufactures hardware, stores every AI conversation by default, or can make third-party providers erase hidden state. ### Narrative spine 1. AI is moving from one-off sessions to long-running relationships with people, teams, tools, and agents. 2. Those relationships need memory, but today memory is fragmented across providers, apps, vector stores, agent logs, and opaque runtime state. 3. Users cannot reliably carry context between AI subscriptions. Developers rebuild memory for every agent. Enterprises struggle to prove what was remembered, retrieved, denied, exported, or tombstoned. 4. Enigma makes AI memory portable and accountable through local-first vaults, Memory Passports, scoped retrieval, boundary-aware gateways, and signed receipts. 5. The long-term company is not another chatbot, model gateway, or vector database. It is the neutral memory/proof substrate beneath AI applications. ## Messaging pillars ### Pillar 1 — Portability Message: Your AI memory should move with you. Support: - Memory Passports carry Enigma-managed context across supported clients and workflows. - Provider-native memory becomes cache, not the durable system of record. - Importers preserve source caveats; imported memory becomes canonical only after Enigma writes it into a vault and emits Enigma receipts. Proof point: Connector profiles support Claude Desktop, Cursor, Kimi Code, VS Code/Cline, Roo Code, OpenCode, and generic MCP. ### Pillar 2 — Custody Message: Keep durable AI memory local-first or customer-controlled. Support: - The local path uses a vault bundle on the user’s filesystem. - Enterprise deployments can evaluate hosted, BYOC, VPC, on-prem, or air-gapped patterns. - Relays and witnesses should handle opaque encrypted records and commitments, not raw memory plaintext. Proof point: The local demo creates a vault, writes memory, creates a context pack, exports a proof bundle, and verifies it using CLI commands. ### Pillar 3 — Proof Message: Receipts make AI memory events inspectable. Support: - Enigma emits signed receipts for declared Enigma-mediated memory lifecycle and boundary events. - Receipt bundles can verify offline without trusting Enigma servers or model providers. - Gateway decisions can produce policy and boundary evidence for enterprise review. Boundary: Receipts prove declared Enigma-mediated operations were signed, ordered, and committed under a stated policy boundary. They do not prove factual truth, model intent, uninstrumented side-channel absence, complete provider deletion, or deletion from external systems unless those systems provide independent evidence. ### Pillar 4 — Provider neutrality Message: Use memory across models instead of surrendering it to one model provider. Support: - Enigma works beneath AI tools through MCP, CLI, connectors, importers, browser extension scaffolds, desktop scaffolds, relay, and gateway surfaces. - It can wrap existing memory systems and model gateways rather than replacing all of them. - Scoped context can be injected into model calls without making the provider the canonical memory store. ### Pillar 5 — Optional network coordination Message: Local Enigma works without a token; optional network roles coordinate shared infrastructure. Support: - Relays route or store opaque encrypted capsule records. - Witnesses attest compact commitments, receipt roots, and checkpoint context. - Gateways provide policy-mediated access and submission paths. - Utility and governance plans require legal review before publication. Required token boundary: The Enigma token, if launched, is intended for utility, governance, and network access within optional relay, witness, and gateway infrastructure; it is not equity, not a revenue share, not a claim on company assets or user data, and not marketed with any expectation of profit. ## Taglines ### Primary taglines - The memory card for AI. - Portable memory. Verifiable use. - Carry context. Prove boundaries. - Memory that moves. Receipts that prove. - Your AI memory, wherever you work. - One Memory Passport. Every AI workspace. ### Technical taglines - Provider-neutral memory and proof for every AI runtime. - Signed receipts for AI memory lifecycle events. - Offline-verifiable evidence for AI memory boundaries. - Scoped context capsules for model and tool calls. - Memory custody beneath agents, models, and gateways. ### Enterprise taglines - Customer-controlled AI memory with verifiable boundaries. - Durable memory without surrendering the evidence trail. - The auditable memory control plane for enterprise AI. - Keep memory under policy. Send only scoped context. - Bring AI memory to the security review. ### Community and network taglines - Build the proof network for AI memory. - Useful infrastructure, not token hype. - Relay encrypted capsules. Witness opaque roots. Verify receipts. - Governance for protocol parameters, not private user memory. ## Voice and tone ### Brand voice - Plainspoken infrastructure confidence. - Precise about proof boundaries. - Helpful to users and developers. - Calm around token topics. - Opinionated about portability and custody without attacking providers. - Trustworthy enough for enterprise legal and security readers. ### Voice rules Do: - Use concrete mechanisms: vault, passport, scoped context, receipt, verifier, gateway, relay, witness. - Say exactly what is verified. - Pair broad trust language with a boundary sentence. - Treat Solana as optional network coordination infrastructure, not the center of the company story. - Say local-first when describing the local product path. - Say customer-controlled when describing enterprise custody modes. - Distinguish personal owner, tenant owner, and delegated controller modes. Do not: - Use hype, scarcity, or financial upside language. - Claim Enigma eliminates all trust, privacy, legal, or operational risk. - Claim deletion from Enigma proves deletion from providers, model weights, backups, logs, screenshots, or people. - Claim complete side-channel absence. - Claim providers are enemies. Enigma is the neutral layer that can wrap, import, export, and verify around them. - Make blockchain sound required for local Enigma. ### Word bank Preferred: - Provider-neutral - Local-first - Customer-controlled - Memory Passport - Scoped context - Boundary receipt - Offline-verifiable - Enigma-mediated event - Opaque checkpoint - Encrypted relay record - Witness quorum - Gateway decision - Utility - Governance participation - Network access - Active operator bonding Use carefully: - Ownership: use only when the owner/controller is clear. - Delete: prefer tombstone, remove from Enigma active serving, or delete-request unless referring to a specific command or system boundary. - Proof: specify proof of what. - Decentralized: specify relay, witness, gateway, or governance function. Avoid: - No leaks - Never leaves your hands - Delete everywhere forever - No server ever - Fully private by default across all modes - Guaranteed compliant - Trustless in all respects - Own a piece of Enigma - Guaranteed compensation - Financial-upside framing - Trading hype - Scarcity claims - Resale speculation - Company-ownership framing ## Approved claims ### Product claims approved for launch copy - Enigma is provider-neutral AI memory/proof infrastructure. - Enigma gives users and organizations portable Memory Passports for AI context. - Enigma is local-first for the local path and supports customer-controlled deployment modes for enterprise evaluation. - Enigma emits signed receipts for declared Enigma-mediated memory lifecycle and boundary events. - Receipts can be verified offline without trusting Enigma servers or AI providers. - Enigma can inject scoped context into provider, model, and tool calls instead of making providers the durable memory system of record. - Enigma supports MCP installability for developers and agents. - Enigma can wrap, import, and export with existing memory systems rather than replacing every stack. - Relay, witness, and gateway infrastructure is planned or available by mode for encrypted relay, opaque checkpoint witnessing, and network access. - Raw memories should stay encrypted/off-chain in network paths. - Enigma provides evidence and controls that support audits, deletion workflows, residency policies, and governance. - The gateway evaluates Enigma enterprise policy and signs decisions; it does not call model providers. ### Proof claims approved with boundary language - Enigma can verify that a memory create, retrieval, context-pack, update, export, tombstone, or delete-request event produced a valid receipt inside Enigma’s boundary. - Enigma can verify that a local proof bundle is internally consistent under the verifier checks. - Enigma can verify that a gateway decision followed a stated policy hash when the gateway receipt is valid. - Enigma can verify that a relay stored an opaque encrypted record or signed a witness checkpoint when the relevant receipt and checkpoint verify. - Enigma can support audit evidence, but it does not guarantee compliance outcomes. Required boundary sentence: Receipts prove that declared Enigma-mediated operations were signed, ordered, and committed under a stated policy boundary; they do not prove factual truth, model intent, complete provider deletion, model forgetting, semantic forgetting, or complete absence of uninstrumented side channels. ### Token and network claims approved after legal review - The planned token is intended for utility, governance, and network access within optional relay, witness, and gateway infrastructure. - Token ownership is not equity, not a revenue share, not a claim on company assets or user data, and not marketed with any expectation of profit. - Local-only Enigma memory, MCP use, and offline verification should work without a token. - Operators may bond tokens to qualify for active relay, witness, or gateway work under published network rules after legal and technical review. - Operator compensation, if any, should be described as service fees or subsidies for verified work, not payment for mere holding or guaranteed compensation. - Governance may cover bounded protocol parameters such as schema upgrades, witness admission criteria, fee schedules, grants, verifier requirements, and operator rules. - Governance must not control private user memories, company equity, dividends, employee decisions, customer contracts, or historical receipt truth. ## Banned claims ### Product and proof claims to ban - Enigma proves a memory is factually true. - Enigma proves what a model intended or reasoned internally. - Enigma proves a model provider deleted all internal copies. - Enigma proves model weights forgot. - Enigma proves provider-native memory disappeared unless the provider supplies independent evidence. - Enigma proves semantic paraphrases are absent. - Enigma proves complete side-channel absence. - Enigma makes AI fully compliant. - Enigma removes all trust, privacy, legal, or operational risk. - Enigma puts all AI memory on-chain. - Enigma requires blockchain for local use. - Enigma replaces every vector database, memory engine, model gateway, or observability tool. ### Token claims to ban - Token ownership is stock, equity, revenue share, dividend right, profit right, or company ownership. - Token holders own Enigma, company assets, customer contracts, or user data. - The token has guaranteed demand, compensation, control, access, or user-data rights. - Holding or locking tokens by itself creates compensation. - Token purchase is required for local memory, MCP use, or offline receipt verification. - Governance can access private memories or rewrite valid historical receipts. - Any financial-upside, resale, trading, scarcity, listing, burn-value, or company-ownership language. ## Audience-specific messaging ### AI power users and creators Job: Carry preferences, projects, style, and context across AI subscriptions and local agents. Headline: Your AI memory should move with you. Message: Enigma keeps your durable AI context in a portable Memory Passport so you can use more than one model or client without starting from zero. Proof point: Install locally, create a vault, connect an MCP-capable client, write memory, and verify a receipt. Primary CTA: Install Enigma locally Objection: I already have chat history. Answer: Chat history is useful, but it is usually trapped in one provider or app. Enigma gives you a separate memory layer with receipts for Enigma-managed events. ### Developers and agent builders Job: Add durable memory, scoped retrieval, and proof receipts without building custom governance infrastructure. Headline: One memory API. Every model. Receipts included. Message: Enigma gives agents MCP-installable memory, local vaults, context packs, importers, connectors, and offline-verifiable receipts. Proof point: The MCP server exposes `enigma_init`, `enigma_remember`, `enigma_search`, `enigma_context_pack`, `enigma_delete`, and `enigma_verify_receipts`. Primary CTA: Run the MCP example Objection: We already use a vector store. Answer: Keep it where it works. Enigma is not just retrieval; it is custody, portability, policy, and receipts around memory lifecycle and boundary events. ### Enterprise AI, security, legal, and compliance Job: Use persistent AI memory while proving retention, policy, deletion workflow, region, key, provider, and boundary decisions. Headline: Durable memory under customer control. Message: Enigma lets enterprises keep AI memory in customer-controlled environments, send providers only scoped context, and produce receipts for Enigma-mediated operations. Proof point: Gateway decisions can include provider, model, region, purpose, sensitivity, policy hash, and memory address metadata while avoiding plaintext-heavy evidence exports. Primary CTA: Request a private pilot Objection: Does this make us compliant? Answer: Enigma supports governance and audit work with evidence. It does not guarantee compliance outcomes; counsel and internal controls decide the compliance posture. ### Memory engines, model gateways, and ecosystem partners Job: Offer memory portability and proof without rebuilding receipt infrastructure. Headline: Make your memory proof-carrying. Message: Enigma can wrap, import, export, and verify around existing memory systems and gateways so partner products can support Memory Passport workflows. Proof point: Receipt schemas, verifier outputs, import reports, capsules, connector profiles, gateway records, and relay records can become integration contracts. Primary CTA: Build an Enigma-compatible integration Objection: Are you replacing us? Answer: No. Enigma is a neutral memory/proof layer. It can complement existing memory engines, observability stacks, gateways, and provider-native features. ### Community and network operators Job: Participate in useful infrastructure without confusing token participation with financial speculation. Headline: Help operate the proof network for AI memory. Message: Enigma’s planned network roles include relay, witness, and gateway infrastructure for encrypted records, opaque checkpoints, and access coordination. Proof point: Relays should reject plaintext-looking memory fields and handle opaque encrypted records. Witnesses should attest commitments and roots, not user memory plaintext. Primary CTA: Read the network roles guide Objection: Does token ownership provide company rights? Answer: No. The planned token is intended for utility, governance, and network access after legal review. It is not equity, not a revenue share, not a claim on company assets or user data, and not marketed with any expectation of profit. ## Comparison angles ### Enigma vs provider-native memory Position: Provider-native memory is convenient; Enigma is portable and proof-carrying. Use when: Explaining lock-in and cross-provider continuity. Copy: Provider memory can improve one app. Enigma is designed to carry Enigma-managed context across apps and produce receipts for lifecycle events inside its boundary. Avoid: Saying provider memory is bad or unsafe by default. ### Enigma vs chat history Position: Chat history is a record; Enigma is a memory layer. Use when: Users say they already save conversations. Copy: Chat history helps you look back. Enigma helps agents retrieve scoped context forward and gives you receipts for Enigma-mediated memory events. Avoid: Claiming Enigma imports every chat perfectly or completely. ### Enigma vs vector databases Position: Vector databases help retrieve; Enigma governs memory custody and proof. Use when: Developers compare Enigma to retrieval infrastructure. Copy: A vector database can find similar text. Enigma manages vault state, Memory Passports, context packs, lifecycle receipts, and boundary evidence. Avoid: Claiming Enigma replaces all retrieval systems. ### Enigma vs agent observability Position: Observability shows traces; Enigma emits portable memory receipts. Use when: Security and developer buyers compare logs to proof bundles. Copy: Traces and dashboards help debug. Enigma receipts are signed evidence for declared memory and boundary events that can be verified outside the dashboard. Avoid: Saying observability tools are not useful. ### Enigma vs model gateways Position: Gateways route calls; Enigma controls and proves memory context. Use when: Enterprise buyers already have a gateway. Copy: A model gateway can enforce provider routing and policy. Enigma can sit beside or in front of gateway workflows to manage durable memory and produce receipts for memory-specific decisions. Avoid: Claiming Enigma calls model providers when the current gateway does not. ### Enigma vs blockchain storage Position: Enigma is not putting memories on-chain. Use when: Crypto-native or privacy-sensitive audiences ask about Solana. Copy: Enigma’s optional network may use Solana for service coordination, opaque commitments, operator records, and governance. Raw memories, prompts, embeddings, transcripts, and private metadata should stay encrypted/off-chain by default. Avoid: Saying blockchain is required for local Enigma or that on-chain commitments eliminate all risk. ## Objection handling ### “Why not just use the AI provider’s built-in memory?” Built-in memory is useful inside one provider. Enigma is for people and teams who want a provider-neutral memory layer, portable Memory Passports, and receipts for Enigma-mediated memory events. ### “Does this prove the model used the memory correctly?” No. Enigma can prove that scoped context was produced or a boundary event was recorded inside Enigma’s system. It cannot prove the model’s internal reasoning, intent, or exact use of context unless a separate instrumented system provides that evidence. ### “Does deletion really delete everything?” Enigma can remove or tombstone memory from Enigma active serving state and produce receipts for that Enigma-controlled transition. It cannot prove deletion from external providers, hidden logs, backups, screenshots, model weights, or human memory. ### “Is this only for crypto users?” No. Local Enigma memory, MCP use, and offline verification are product-first paths. The planned token is for optional network utility, governance, and access around relay, witness, and gateway infrastructure, subject to legal review. ### “Will enterprises need token custody?” Recommended draft value requiring board/legal review: no. Enterprise customers should be able to buy product and gateway access through normal billing while gateways handle any required network settlement where legally and operationally appropriate. ### “Why Solana?” Solana is a high-throughput public network with established token primitives. If Enigma uses Solana, it should be for optional network coordination such as operator records, service settlement, governance, and opaque checkpoint anchoring. The product must not require Solana for local memory or offline receipt verification. ### “SPL Token or Token-2022?” Recommended draft value requiring board/legal review: use the simplest compatible token design unless a Token-2022 extension is needed at launch. SPL tokens use Mint Accounts and Token Accounts. Token-2022 adds optional extensions; most extensions must be planned at mint creation. MetadataPointer and TokenMetadata can store name, symbol, URI, update authority, and custom metadata on the mint. Final choice requires engineering, wallet-compatibility, and legal review. ### “Can governance take my data?” No official messaging should grant governance control over private user memories. Governance, if launched, should be bounded to protocol parameters such as fee bands, operator rules, schema upgrades, witness criteria, grants, verifier requirements, and treasury/service budgets. ### “Are receipts enough for audits?” Receipts are evidence, not a complete audit program. They can support audits by showing Enigma-mediated operations under stated policies. Enterprise controls, legal review, data processing terms, identity systems, retention policy, and operational processes remain necessary. ## Naming and capitalization - Enigma - Memory Passport - Memory Capsule - Receipt - Verifier - Vault - Gateway - Relay - Witness - Checkpoint - Tombstone - Scoped context - Boundary receipt - Enigma-mediated event - MCP server - Local-first - Customer-controlled ## Messaging hierarchy by surface ### Homepage Lead with: The memory card for AI. Support with: Provider-neutral memory, local-first custody, scoped context, signed receipts, MCP installability. Boundary: Receipts verify Enigma-mediated operations, not provider internals or model truth. CTA: Install locally, verify a receipt, request enterprise pilot. ### Developer docs Lead with: One memory API, every model, receipts included. Support with: MCP tools, connector profiles, CLI commands, receipt schema, verifier, importers, gateway and relay demos. Boundary: Receipts verify Enigma operations, not model reasoning. CTA: Run `enigma-mcp` and verify a proof bundle. ### Enterprise collateral Lead with: Durable AI memory under customer control. Support with: gateway policy decisions, BYOC/VPC/on-prem modes, KMS/BYOK posture where configured, minimized SIEM evidence, audit support. Boundary: Supports governance and audits; does not guarantee compliance or external-provider deletion. CTA: Run a private pilot. ### Network and community Lead with: Useful infrastructure for AI memory proof. Support with: relay, witness, gateway, opaque commitments, service access, bounded governance. Boundary: Token is utility/governance/network-access oriented and legally reviewed; no financial-upside framing. CTA: Learn network roles and join the community. ## Recommended launch message set ### X / Farcaster short post AI memory should not be trapped in one app. Enigma is the memory card for AI: portable Memory Passports, local-first custody, scoped context, and offline-verifiable receipts for Enigma-mediated memory events. Install locally. Connect MCP. Verify a receipt. ### LinkedIn launch post AI teams are moving from one-off prompts to durable agent workflows. That shift creates a new infrastructure problem: memory has to be portable, governed, and verifiable. Enigma is provider-neutral AI memory/proof infrastructure. It gives users and organizations a local-first or customer-controlled vault, Memory Passports for portable context, scoped context injection into model and tool calls, and signed receipts for Enigma-mediated memory lifecycle and boundary events. The goal is simple: make AI memory work more like trusted portable storage than a feature trapped inside one subscription. ### Enterprise email opening Your AI pilots are beginning to remember. The question is whether your organization can control that memory, scope what crosses model boundaries, and verify the evidence later. Enigma provides a provider-neutral AI memory control plane with customer-controlled deployment paths, scoped retrieval, gateway policy decisions, minimized evidence exports, and offline-verifiable receipts for Enigma-mediated events. ### Developer email opening If your agent needs durable memory, you should not have to build vault custody, import/export, policy boundaries, and receipt verification from scratch. Enigma gives you a provider-neutral memory/proof layer through MCP, local bundles, context packs, and offline-verifiable receipts.