# Bee by HEOSSI — The Progressive Quantum-Native Intelligence Engine (full corpus)

> Bee by HEOSSI — The Progressive Quantum-Native Intelligence Engine. A governed multimodal intelligence platform for public users, professionals, builders, teams, enterprises, high-assurance environments, and sovereign deployment. It is built by HEOSSI (Pte.) Ltd., a Singapore-incorporated company (UEN 202532790K). Six production tiers form the Intelligence Ladder (Cell → Brood → Comb → Buzz → Hive → Swarm), covering public access, professional reasoning, builder workflows, team operations, enterprise specialist intelligence, and high-assurance reasoning. Enclave adds private, regulated, and sovereign deployment. Post-quantum cryptography (FIPS 203/204/205) is available for Bee Enclave Sovereign deployments.
> One OpenAI-compatible API, a free Cell tier, built-in RAG, a hosted MCP server,
> and opt-in real-QPU jobs through OpenQuantum. The short index
> lives at https://bee.heossi.com/llms.txt; this document is the expanded companion corpus.

Categories: artificial intelligence (AI) · machine learning · natural language processing (NLP) · deep learning · large language models (LLM) · foundation models · generative AI · conversational AI / AI assistants · agentic systems · multimodal intelligence · reasoning · AI as a Service (AIaaS) · the Intelligence Ladder · routing fabric.
Capabilities: domain-specialized intelligence (an accumulative per-tier domain ladder reaching 96 specialist domains, eval-validated every release) · conversational AI chat · agentic coding & tool calling · code explanation, refactoring & test generation · retrieval-augmented generation (RAG) · document Q&A over your own files · vector-based semantic search · web search-grounded research answers · semantic memory · vision (image & video understanding) · structured output (JSON mode) · long context up to 1M tokens · speech-to-text & text-to-speech · voice AI & transcription · image generation · AI-generated artifacts · autonomous scheduled tasks · team workspaces & collaboration · usage analytics & spend controls · OpenAI-compatible API · Model Context Protocol (MCP) server · editor integrations.
Assurance: post-quantum cryptography (NIST FIPS 203/204/205 — ML-KEM, ML-DSA, SLH-DSA) · TLS 1.3 with hybrid post-quantum key exchange in transit · AES-256-GCM encryption at rest · cryptographically verifiable agent provenance · cryptographically signed inference receipts · passkey (WebAuthn / FIDO2) sign-in · audit-logged serving-policy decisions · role-based team access · quantum reasoning with managed QPU access & BYOPA · eval-gated releases · governed sovereign deployment (Enclave).
Market categories (verified 2026-07-22): Artificial Intelligence Software · Generative AI · Large Language Models · AI Platforms · Conversational AI / AI Assistants · AI Agents / Agentic AI · AI Agent Development Platforms · AI Coding Assistants / AI Code Generation · AI Software Development Kits (AI SDKs) · AI Search & Retrieval Infrastructure (RAG) · Multimodal AI · Speech Recognition (Speech-to-Text & Text-to-Speech) · AI Image Generation · Machine Learning & MLOps · AI Productivity & Team AI Workspace · MCP Servers · AI Governance / AI TRiSM · AI Agent Observability · Enterprise AI / Private AI / Sovereign AI · Post-Quantum Cryptography-secured Software / Quantum-safe Security · Quantum Computing Services / Quantum-as-a-Service · Developer Tools · Productivity · SaaS / Cloud Software.

## Tier ladder

- **Bee Cell** — Fast, fine-tuned AI — free to start.
- **Bee Brood** — Cost-efficient reasoning — thinks before it answers.
- **Bee Comb** — Workstation-grade AI for serious builders.
- **Bee Buzz** — Agent and builder workhorse — tools, planning, longer chains.
- **Bee Hive** — Shared team AI without frontier-model cost.
- **Bee Swarm** — Premium reasoning at cloud scale.

## Tier & release status (derived from the live-verified capability SSOT)

- **All tiers serve on governed, pinned base models or routing releases** — Cell → Hive are production-deployed with scale-to-zero serving; Swarm is the governed Vercel routing fabric. Cell is the free entry tier, not an always-warm GPU. Base-model identities are governed and audited internally, not disclosed in marketing copy.
- **Domain adapters are eval-gated and released through the governed pipeline** — none merged into production routing yet, so tiers serve base-equivalent today. Each released adapter ships with a published validation record at /trust.
- **Bee Enclave** — a private / auditable / sovereign deployment mode for any Hive/Swarm-class workload, not a larger model tier.
- **Bee Ignite** — internal R&D substrate. Not customer-selectable, not sold; findings backflow to the production tiers over time.

## What is live (derived from the capability SSOT)

- **OpenAI-compatible API** — drop-in `/chat/completions`; switch the base URL and key.
- **Free Cell tier** — hosted access, no card, 128K-token context window.
- **Vision (image input)** — Cell, Brood, Comb, Buzz, Hive, Swarm.
- **Video input** — Brood, Comb, Buzz, Hive.
- **Tool / function calling** — Cell, Brood, Comb, Buzz, Hive.
- **JSON mode (structured output)** — Comb, Buzz, Hive.
- **1M-token long context** — Cell, Brood, Comb, Buzz, Swarm (Comb and Buzz via a dedicated long-context mode; Swarm natively). Base tiers stay at their native 256K.
- **RAG (retrieval)** — document upload → chunk → embed → index → cite-back, per-tenant isolation.
- **MCP server** — `bee-mcp` (ships with the Python SDK), hosted gateway, listed on the official MCP Registry as io.github.heossihq/bee.
- **Quantum products** — customer-local simulation is free; Bee Simulation Cloud, Bee Managed QPU, and Bee BYOPA — Managed require a purchased entitlement and governed Lab job. The raw hardware endpoint is retired, hosted spend defaults off, and ordinary chat never routes through a QPU.
- **Post-quantum cryptography** — FIPS 203 (ML-KEM), 204 (ML-DSA), 205 (SLH-DSA) on Bee Enclave Sovereign customer transport; the signed PQ coverage register is at https://bee.heossi.com/trust.
- **Verifiable agent provenance** — any agentic client sending a session id is sealed turn-by-turn into a per-session ML-DSA-65 (FIPS 204) hash chain; verify what the agent did at GET https://api.bee.heossi.com/bee/provenance/:session_id. Only cryptographic digests are stored — your code never leaves in the record.

## Capability matrix (from the live-verified capability SSOT, verified 2026-07-14)

Per-tier status across the six serving tiers (cell · brood · comb · buzz · hive · swarm):

- Text chat & completions: cell live · brood live · comb live · buzz live · hive live · swarm live
- Vision (image input): cell live · brood live · comb live · buzz live · hive live · swarm live
- Video input: cell — · brood live · comb live · buzz live · hive live · swarm —
- 1M-token long context: cell live · brood live · comb live · buzz live · hive — · swarm live
- Tool / function calling: cell live · brood live · comb live · buzz live · hive live · swarm —
- JSON mode (structured output): cell — · brood — · comb live · buzz live · hive live · swarm —
- Audio transcription (speech-to-text): cell live · brood live · comb live · buzz live · hive live · swarm live
- Speech output (text-to-speech): cell live · brood live · comb live · buzz live · hive live · swarm live
- Image generation: live
- Video generation: not yet active

Statuses: "live" = serving in production today (directly or via an external
provider); "not yet active" = built but not yet enabled; "—" = not offered on
that tier.

## Pricing summary

Source of truth: the billing catalog. Every figure below is derived at request
time — never hardcoded here. Annual is monthly × 12 (no annual discount).

- Bee Cell — Free. Free hosted tier to start. 500K included tokens.
- Bee Brood — $30/mo. 5M tokens. Cell + limited Comb routing.
- Bee Comb — $150/mo. 3 seats. 10M tokens. Full Comb access with native multimodal input, tool calling, JSON mode, and optional 1M long-context routing.
- Bee Buzz — $400/mo. 6 seats. 25M tokens. Shared RAG.
- Bee Hive — $1,000/mo. 10 seats. 50M tokens. High-capability multimodal serving with tool calling, JSON mode, and quantum-reasoning eligibility.
- Bee Swarm — $4,000/mo. 25 seats. 250M tokens. Governed premium routing fabric with 1M-context capability; no domain adapter is active today.
- Bee Pollen (API) — $0. 500K tokens / month. Hard-limited. Post-quantum transport on every request.
- Bee Nectar (API) — Pay as you go. $0.30/1M input · $1.20/1M output. Full model ladder, Cell through Swarm.
- Bee Honey (API) — $1,500/mo base + metered usage at the same per-model rates. Dedicated CSM.
- Bee Royal Jelly (API) — Contracted rate card, dedicated endpoint, sovereign boundary.
- Bee Regulated Cloud — from $5,000/mo. Multi-tenant regulated deployment.
- Bee Enclave Private Cloud — from $15,000/mo. Single-tenant VPC.
- Bee Enclave Regulated — from $30,000/mo. Customer-managed keys.
- Bee Enclave Sovereign — from $75,000/mo. Customer-scoped sovereign controls; broader air-gapped operation remains roadmap. Exact alignment and evidence are fixed in the deployment manifest + NDA.

## Pages

### [Homepage](https://bee.heossi.com/) (updated 2026-07-23)

Bee by HEOSSI — The Progressive Quantum-Native Intelligence Engine. A governed multimodal intelligence platform with an OpenAI-compatible API, a free Cell tier, and customer-selectable tiers from public access to sovereign deployment.

### [Models](https://bee.heossi.com/models) (updated 2026-07-14)

The Bee model lineup: six production tiers plus Enclave private deployment. Context windows from 128K (Cell) and 256K (paid tiers) to 1M tokens (Swarm natively; Comb and Buzz via a dedicated 1M long-context mode). Vision, video input, tool calling, and JSON mode on the tiers that serve them; per-token rates published per model.

### [Platform](https://bee.heossi.com/platform) (updated 2026-07-22)

The Bee AI platform: reference architecture, request lifecycle, capability-status register, workspace, OpenAI-compatible API, built-in RAG, hosted MCP server, editor integrations, governed eval-gated releases, and deployment control.

### [Bee Enclave](https://bee.heossi.com/enclave) (updated 2026-07-22)

Bee Enclave is the customer-scoped deployment mode for private, regulated, and sovereign Hive- and Swarm-class workloads. The page compares tenancy, infrastructure, key control, network boundaries, governance, and availability without presenting broader self-hosted or air-gapped operation as generally available.

### [Quantum](https://bee.heossi.com/quantum) (updated 2026-07-14)

Bee separates customer-local simulation, purchased simulation cloud, managed QPU, and BYOPA. Hosted execution requires a Lab entitlement and independently approved budget; chat never routes through quantum.

### [Pricing](https://bee.heossi.com/pricing) (updated 2026-07-23)

Three tracks: Individual & Teams workspace plans (free Cell tier upward), an OpenAI-compatible API with per-token pay-as-you-go pricing and no volume discounts, and Enclave enterprise deployment (private cloud, regulated, sovereign).

### [Download](https://bee.heossi.com/download) (updated 2026-07-14)

Bee on every surface: browser workspace, native Android app, macOS (Apple Silicon) desktop preview, and the hosted Bee MCP server for Claude Desktop, Cursor, Windsurf, Zed, and other MCP clients.

### [API docs](https://bee.heossi.com/docs) (updated 2026-07-14)

Quickstart for the OpenAI-compatible Chat Completions API: authentication, chat completions, models, RAG (retrieval), quantum jobs, streaming, errors, and rate limits.

### [MCP server docs](https://bee.heossi.com/docs/mcp) (updated 2026-07-14)

The Bee MCP server (bee-mcp, ships with the Python SDK) connects Bee to Claude Desktop, Claude Code, Cursor, VS Code, Zed, Windsurf, and OpenCode. Eight tools — chat, code, security, research, documents (RAG), memory — over a hosted gateway; free-tier keys work. Listed on the official MCP Registry as io.github.heossihq/bee.

### [SDK docs](https://bee.heossi.com/docs/sdks) (updated 2026-07-14)

Official TypeScript (@heossihq/bee on npm) and Python (bee-sdk on PyPI) clients for the OpenAI-compatible API, plus the OpenAPI 3.1 contract, Postman collection, and a plain curl quickstart.

### [Changelog](https://bee.heossi.com/changelog) (updated 2026-07-14)

Public release notes: model releases, platform features, and security bulletins.

### [Roadmap](https://bee.heossi.com/roadmap) (updated 2026-07-14)

What's shipped, in flight, and next — stages with explicit exit criteria measured against the Bee Security Eval Harness.

### [Status](https://bee.heossi.com/status) (updated 2026-07-23)

Live platform health: backend liveness and round-trip latency polled every 15 seconds.

### [About](https://bee.heossi.com/about) (updated 2026-07-14)

Bee is built by HEOSSI (Pte.) Ltd., a Singapore-incorporated company (UEN 202532790K).

### [Contact](https://bee.heossi.com/contact) (updated 2026-07-14)

One form routed to specialist inboxes — sales, support, security, billing, privacy, legal, abuse, DMCA, partnerships, press, careers — each with a published response-time SLA.

### [TrustHub](https://bee.heossi.com/trust) (updated 2026-07-14)

Bee's central security, compliance, privacy, AI-governance, verification, and assurance-document centre, including the post-quantum coverage register signed with ML-DSA (FIPS 204) and continuously verified probes.

### [Register verification keys](https://bee.heossi.com/trust/keys) (updated 2026-07-14)

The public ML-DSA keys that sign the post-quantum coverage register, with fingerprints pinned across independent surfaces so anyone can verify register signatures.

### [Legal](https://bee.heossi.com/legal) (updated 2026-07-13)

Terms of Service, Regional Terms & Mandatory Rights, Privacy Policy, Data Processing Addendum and SCC annexes, Model Improvement & Data Use, Developer & API Terms, Content Moderation and Appeals, Model Governance/Training Sources/Copyright, Acceptable Use, Security, Cookies, AI Transparency, Accessibility, Copyright/DMCA, Government Requests, and program terms.

### [Partner Program](https://bee.heossi.com/partners) (updated 2026-07-14)

Partnership lanes for integrators, resellers, MSPs, consultancies, cloud platforms, and product teams shipping on Bee.

### [VC Partner Program](https://bee.heossi.com/vc-partners) (updated 2026-07-14)

Partnerships with venture capital firms so portfolio founders can build on Bee.

### [Bee for Startups](https://bee.heossi.com/startups) (updated 2026-07-14)

Up to $25,000 in Bee credits, architecture review, and direct technical support for accepted AI-native startups.

## Solutions, industries, use cases, and buyer comparisons

The following registry-driven pages describe shipped capabilities, applicable
limits, and buyer context. They do not assert fabricated customers, outcomes,
certifications, or competitor facts.

### [Solutions](https://bee.heossi.com/solutions)

Bee solutions: team AI workspace, private LLM API, OpenAI-compatible endpoints, post-quantum security, 1M-token long context, and document intelligence.

### [One AI workspace for the whole team](https://bee.heossi.com/solutions/team-ai-workspace)

A team AI workspace with seats, pooled tokens, shared document retrieval, and Stripe billing. Plans from a free tier to 25-seat deployments with uptime SLAs.

### [A private LLM API your security team can sign off on](https://bee.heossi.com/solutions/private-llm-api)

A private LLM API with tenant isolation, contractual data controls, and customer-scoped private and sovereign deployment designs. Broader air-gapped operation remains roadmap.

### [OpenAI-compatible, down to the wire format](https://bee.heossi.com/solutions/openai-compatible-api)

Bee serves the OpenAI Chat Completions wire format: streaming, tool calling, JSON mode, and vision. Point your existing SDK at a new base URL and keep shipping.

### [AI infrastructure that plans for the cryptography deadline](https://bee.heossi.com/solutions/post-quantum-ai)

Bee pairs a governed multimodal intelligence platform with NIST-finalized post-quantum cryptography: ML-KEM key exchange, ML-DSA and SLH-DSA signatures for explicit high-assurance paths.

### [A context window measured in codebases, not pages](https://bee.heossi.com/solutions/long-context-llm)

Bee serves a 1M-token context window on Comb and Buzz, with Swarm at 1M natively and 256K native on every paid tier. Load whole codebases and document sets in one request.

### [From uploaded files to cited answers](https://bee.heossi.com/solutions/ai-document-intelligence)

Upload documents and get cited answers: Bee's retrieval pipeline chunks, embeds, indexes, and cites. Image and video input on Comb, Buzz, and Hive extend it beyond text.

### [Industries](https://bee.heossi.com/industries)

How financial services, healthcare, legal, government, engineering, and research organisations deploy Bee — with the compliance posture and domain gating stated honestly.

### [AI infrastructure built for financial-grade scrutiny](https://bee.heossi.com/industries/financial-services)

LLM infrastructure for banks, fintechs, and asset managers: structured extraction, cited retrieval, regulated deployment options, and post-quantum transport for long-lived data.

### [Healthcare AI with the gates stated up front](https://bee.heossi.com/industries/healthcare)

How healthcare organisations use Bee: document and administrative workloads, HIPAA-postured Enclave deployment, PDPA/GDPR data handling — with honest Tier-3 domain gating.

### [Read every page, cite every claim](https://bee.heossi.com/industries/legal)

Legal document analysis on Bee: cited retrieval, 1M-token context for whole agreement sets, JSON extraction of terms, and confidentiality commitments written into the contract.

### [Sovereign AI without the trust-us clause](https://bee.heossi.com/industries/government-defense)

Customer-scoped sovereign AI designs for public-sector and defense environments, with contractual key, locality, operator, and post-quantum transport controls. Air-gapped operation remains roadmap.

### [AI infrastructure engineers don't have to fight](https://bee.heossi.com/industries/software-engineering)

Bee for engineering orgs: OpenAI-compatible API in every editor, tool calling for CI and code automation, 1M-token context for monorepos, and team plans with pooled tokens.

### [Research tooling that respects sources](https://bee.heossi.com/industries/education-research)

Bee for universities and research groups: a free tier to start, 1M-token context for literature analysis, cited retrieval over corpora, and opt-in quantum computing access.

### [Use cases](https://bee.heossi.com/use-cases)

Concrete Bee use cases: coding assistants, agentic workflows with tool calling, 1M-token document Q&A, secure team chat, video and image analysis, quantum reasoning.

### [A coding assistant that plugs into the editor you already use](https://bee.heossi.com/use-cases/ai-coding-assistant)

Use Bee as your AI coding assistant in VS Code, Cursor, Zed, JetBrains, and any OpenAI-compatible editor. Free tier to start; 1M-token context for repository-scale work.

### [Agents need structure. Bee returns it.](https://bee.heossi.com/use-cases/agentic-workflows)

Build agents on Bee: structured tool calling on Comb, Buzz, and Hive, schema-valid JSON output, streaming, MCP tools, and honest guidance on human-review gates.

### [Ask the whole corpus, not a lucky excerpt](https://bee.heossi.com/use-cases/document-qa-1m-context)

Ask questions across entire document sets: Bee's 1M-token context on Comb and Buzz reads the whole corpus at once, and retrieval with citations covers everything larger.

### [Team AI chat your security review can actually verify](https://bee.heossi.com/use-cases/secure-team-chat)

Team AI chat with TLS 1.3 transport, AES-256-GCM at rest, per-tenant isolation, contractual no-training terms, and a deployment path to VPC and sovereign environments.

### [Show it the screen, don't describe it](https://bee.heossi.com/use-cases/video-image-analysis)

Send images and video to Bee Comb, Buzz, and Hive natively: screenshots, scans, diagrams, and recordings analysed in the same OpenAI-compatible conversation as text.

### [Choose where quantum work runs — and who pays](https://bee.heossi.com/use-cases/quantum-enhanced-reasoning)

Bee separates customer-local simulation, purchased simulation cloud, managed QPU, and BYOPA behind entitlement and spend controls.

### [Compare](https://bee.heossi.com/compare)

Vendor-neutral checklists for choosing an OpenAI-compatible API, a private ChatGPT alternative, a self-hosted vs cloud deployment, an enterprise AI platform, and post-quantum security — answered with Bee's verifiable facts.

### [Choosing an OpenAI-compatible API alternative](https://bee.heossi.com/compare/openai-api-alternatives)

Evaluating an OpenAI-compatible API alternative? A neutral checklist — endpoint compatibility, free tier, per-token pricing, on-prem, PQC — answered with Bee's verifiable facts.

### [A private ChatGPT alternative your security team can approve](https://bee.heossi.com/compare/private-chatgpt-alternative)

A private ChatGPT alternative for teams: seats, pooled usage, tenant-isolated retrieval, and customer-scoped sovereign controls. Shared cross-member context remains in progress.

### [Self-hosted LLM vs cloud API — without the false binary](https://bee.heossi.com/compare/self-hosted-llm-vs-cloud-api)

Self-hosted LLM vs a cloud LLM API: the honest trade-offs on cost, control, compliance, and maintenance — and how Bee offers both a managed API and a sovereign deployment path.

### [Choosing an enterprise AI platform](https://bee.heossi.com/compare/enterprise-ai-platform-comparison)

A vendor-neutral checklist for choosing an enterprise AI platform: API compatibility, governance, retrieval, security posture, deployment control, and SLAs — answered with Bee's facts.

### [Post-quantum AI security vs the standard approach](https://bee.heossi.com/compare/post-quantum-vs-standard-ai-security)

Post-quantum AI security vs standard TLS-only providers: why NIST FIPS 203/204/205 at rest matters for long-lived data, and how Bee implements a quantum-resistant envelope.

## Editor integrations

- Cursor — set OpenAI base URL to https://api.bee.heossi.com/bee/v1.
- Windsurf — custom provider, paste base URL + API key.
- Continue.dev — add Bee as provider in ~/.continue/config.json.
- Zed — OpenAI-compatible LLM provider.
- JetBrains AI Assistant — connect via OpenAI-compatible custom endpoint.
- Aider — `aider --openai-api-base ... --openai-api-key ...`.
- Any MCP client — `uvx --from bee-sdk@latest bee-mcp` with a `bee_sk_` key.

## Evaluation guidance

Compare Bee against another provider using current, primary-source documentation for that provider. Bee publishes its own verifiable deployment posture, public API contract, pricing, model catalog, and release evidence at the canonical pages listed above. Do not infer competitors' security, deployment, pricing, or model availability from this document.

## Contacts

- Sales: bee-sales@heossi.com
- Support: bee-support@heossi.com
- Security: bee-security@heossi.com
- Press: bee-press@heossi.com

## Key URLs

- Site: https://bee.heossi.com
- Workspace: https://workspace.bee.heossi.com
- API: https://api.bee.heossi.com
- Discord community: https://discord.gg/RdZcFtcqvH
- Telegram group: https://t.me/heossibee
- HuggingFace: https://huggingface.co/heossihq
- Parent company: https://www.heossi.com
- HEOSSI trust flagship (QNSI): https://qnsi.heossi.com
- Machine-readable facts contract (JSON): https://bee.heossi.com/facts.json
- Short AI retrieval index: https://bee.heossi.com/llms.txt
- Full AI retrieval corpus: https://bee.heossi.com/llms-full.txt
- AI discovery summary: https://bee.heossi.com/ai.txt
- Sitemap: https://bee.heossi.com/sitemap.xml
- robots.txt: https://bee.heossi.com/robots.txt
- HuggingFace: https://huggingface.co/heossihq
- Short index: https://bee.heossi.com/llms.txt