AlphaBitCore vs OpenBox AI (2026)

Agents take authenticated action in outside systems through a governed tool layer. Every model, agent, tool and workflow is invoked through one SDK and HTTP contract. The tool registry scopes every MCP server by authority, and the Workbench ships 10+ MCP connectors including primary financial data integrations. The Effector is the sole component holding durable mutation credentials and commits an external effect only within the scope the Gateway authorized. Beyond the connector catalog, customers can register their own MCP servers, internal APIs and data systems as tools.

SourceAlphaBitCore, alphabitcore.com (platform/unify, developers, home)read 2026-09-25

SDKs and plugins wrap agents built on Temporal (a Python plugin), LangChain, LangGraph, CrewAI and Deep Agents in Python, and Mastra and CopilotKit Runtime v2 in TypeScript. An n8n community node and an OpenRouter SDK round out the set. Each operation the agent takes passes through OpenBox for a verdict. These integrations govern actions the customer's own framework takes. Agents cannot take authenticated action in an outside system through OpenBox. Cursor and OpenClaw integrations are coming soon.

SourceOpenBox, docs.openbox.ai/getting-started and /developer-guideread 2026-09-25

AlphaBitCore ships a multi step, multi agent runtime with a Planner (task decomposition and candidate actions), an Orchestrator, an Agent Runtime and a Skills and Tool Registry, and customers can bring LangGraph, Bedrock or Temporal instead. Workflows compose skills, tools and models into governed pipelines (95 in the Workbench), and Agent Teams run a Team Room that calls member agents and returns one synthesized answer.

SourceAlphaBitCore, alphabitcore.com (developers, home, developers/mcp)read 2026-09-25

The customer's framework, such as Temporal, LangGraph or CrewAI, owns the workflow, its steps and its handoffs, and OpenBox governs those runs without running them. A HALT verdict cancels the framework's workflow. Multi-Agent Sessions is a view of each team's runs. OpenBox names no workflow model of its own.

SourceOpenBox, docs.openbox.ai/core-concepts/governance-decisions and /administration/organization/teams/multi-agent-sessionsread 2026-09-25

Work reaches an agent when a caller sends it, either a user in the Prime app or an external assistant (ChatGPT, Claude, Copilot, Codex) delegating a task through Prime MCP. Those are ways to call an agent, not events that start one, and there is no schedule, webhook or event source. The Market and Earnings Briefer describes daily briefings, but AlphaBitCore does not say what runs them.

SourceAlphaBitCore, alphabitcore.com (developers/mcp, Workbench agents)read 2026-09-25

The customer's own runtime, such as Temporal, LangChain or n8n, starts and runs the agent. OpenBox's SDKs wrap that agent, and every governance evaluation responds to an operation the agent began. Alerts on trust tier changes and goal drift flag agents for people to review, and they do not wake an agent.

SourceOpenBox, docs.openbox.ai/getting-started and /dashboard/alerts; docs.openbox.ai/dashboard/alertsread 2026-09-25

Context is assembled per run from connected sources, not held in a maintained retrieval structure over the customer's corpus. Workbench agents pull entity, financials, holdings and news through workflows and MCP connectors, and the invocation contract can call a retrieval service the customer already runs. There is no index, graph or embeddings layer that AlphaBitCore maintains over the customer's documents, and the Nexus gateway's semantic cache stores responses, not knowledge.

SourceAlphaBitCore, alphabitcore.com (Workbench agents, platform/unify, nexus)read 2026-09-25

OpenBox is a governance layer around the customer's agent, and it names no way to index or retrieve the customer's documents for the agent to ground its answers on. Guardrails inspect the agent's inputs and outputs.

SourceOpenBox, docs.openbox.ai/overview and the docs index (llms.txt and llms-full.txt)read 2026-09-25

Conversation state lasts within a session. A Prime conversation continues across messages through a conversation handle (handles expire after 15 minutes by default). There is no memory layer with a set scope and lifetime. The Context Store in the record plane captures replay inputs for audit, and the Nexus gateway's semantic cache stores responses. Neither is memory the agent reads to decide.

SourceAlphaBitCore, alphabitcore.com (developers/mcp sections 3 and 4, developers)read 2026-09-25

Trust score history, approval history and behavioral rules that track patterns over several steps all describe the agent. OpenBox applies them as governance rules, and the agent never reads them as context for its decisions. The Enterprise plan includes Universal Agent Memory Governance, which governs the customer's memory instead of providing one, and OpenBox describes it no further. OpenBox names no memory layer or state the agent reads across runs.

SourceOpenBox, docs.openbox.ai/core-concepts/trust-scores, /trust-lifecycle/authorize/behaviors and openbox.ai/pricingread 2026-09-25

A person approves before the action commits, and the customer sets what needs sign off. A Prime execution suspends with status HITL_REQUIRED, and the owner approves, rejects or cancels it (from an MCP client or the Prime Console) before it continues. Policy as code at the Gateway evaluates scope, authority and guardrail rules on every call, and a denied execution produces zero state change and a sealed denial. The Workbench adds a tiered supervisory approval queue with role aware routing.

SourceAlphaBitCore, alphabitcore.com (developers/mcp section 4, platform/govern, home)read 2026-09-25

When an operation crosses a risk threshold, a policy calls for review or the agent's trust tier requires it, OpenBox returns REQUIRE_APPROVAL. The operation pauses and appears in the Approvals queue, where a reviewer approves, rejects or escalates it. An unanswered request expires. Who approved or denied, and when, is recorded with the decision.

The Adapt tab shows each agent's pending approvals with risk tier, semantic action type and triggering rule, plus approval history and approval rate. Chained guardrails screen the inputs and outputs of every operation, and BLOCK and HALT verdicts stop an action or end the session. Each framework SDK has its own way of enforcing these verdicts.

SourceOpenBox, docs.openbox.ai/core-concepts/governance-decisions, /trust-lifecycle/adapt, /trust-lifecycle/authorize/guardrails and /administration/compliance-and-auditread 2026-09-25

Access controls are in place, and a compliance audit is still pending. Every call carries a verifiable caller identity through one identity plane, and the Gateway federates with the enterprise identity provider (OIDC, SPIFFE and workload identity, specified per deployment).

Prime MCP signs users in with OAuth 2.0 and PKCE and derives tenant, user and role from the token, scope and authority policies are bound to each agent, and the supported gateway tier offers enterprise SSO and SCIM. The SOC 2 program is in progress, with the audit roadmap and third party attestations shared only in private briefings. Framework mappings are published, but they describe posture, not a completed audit.

SourceAlphaBitCore, alphabitcore.com (trust, platform/unify, developers, developers/mcp, pricing)read 2026-09-25

Admin, Developer and Viewer roles are set under Organization > Permissions. A Viewer can see agents, logs and reports but cannot change anything. Teams decide which agents a member can see and which sessions they can act on. Each agent also carries its own cryptographic identity, a DID and an Ed25519 signing key kept separate from its API key. OPA/Rego policies run stateless permission checks on every agent operation, giving least privilege tool access. An organization audit log records who changed what. OpenBox names no identity provider integration such as SSO, SAML or SCIM.

OpenBox shows ISO 27001, ISO 42001 and NIST badges on its site, but none links to a certificate, auditor, report or date. It names SOC 2, ISO 27001 and GDPR only as frameworks it maps its customers' audit evidence to.

SourceOpenBox, docs.openbox.ai/administration/organization, /administration/organization/teams, /core-concepts/agent-identity and /trust-lifecycle/authorize/policies, plus the openbox.ai footer and /solution/complianceread 2026-09-25

Every governed execution writes a signed, append only, Merkle sealed event stream with each stage (authorized, planned, executed, sealed) and a capability result envelope naming what was invoked and its inputs, so a customer can inspect each run after the fact. Denials are recorded as sealed events too.

An Inspection, Replay and Monitoring plane reads from the stream for audit, anomaly detection and forensics, and any execution can be re-run from its sealed stream. Execution Attestation Certificates are issued under the customer's keys, and the stream exports to SIEM and observability tooling.

SourceAlphaBitCore, alphabitcore.com (platform/prove, developers, home)read 2026-09-25

Session Replay steps through an agent session event by event, showing every tool call, governance decision and full JSON payload. The Monitor tab tracks each agent's invocations, response time and token use, along with cost and errors, over selectable time ranges.

Every governance decision goes into an immutable audit trail with verdict, reason, workflow and run IDs and approval metadata, and the trail exports on demand. Each session's events are hashed into a Merkle tree and signed into a tamper evident proof certificate, through AWS KMS by default or the customer's own attestation service. An organization audit log records user and configuration changes, separate from the runtime traces.

SourceOpenBox, docs.openbox.ai/trust-lifecycle/session-replay, /trust-lifecycle/monitor, /administration/compliance-and-audit and /administration/attestation-and-cryptographic-proofread 2026-09-25

Three deployment shapes are offered, hosted (Prime only), customer VPC and on premises (all three tiers), with air gapped deployment named for sovereign data constraints. The open source Nexus gateway self hosts by default or deploys from AWS Marketplace. Residency and processing region are specified per deployment, and single tenant and multi tenant patterns are both supported.

SourceAlphaBitCore, alphabitcore.com (pricing, trust, nexus)read 2026-09-25

The Enterprise plan includes Custom Deployment and SLAs, with no named environment, region or way to choose one. Every SDK guide points the agent at the hosted endpoint core.openbox.ai. OpenBox publishes no region list and names no self hosted or customer cloud option.

External attestation lets the customer sign proof certificates with its own attestation service or a Trusted Execution Environment. That choice sets where the signing key sits, not where OpenBox runs. OpenRouter routing policies can restrict model providers by data residency, which applies to the customer's model traffic, not to where OpenBox itself is deployed.

SourceOpenBox, openbox.ai/pricing, docs.openbox.ai/getting-started/deep-agents and /administration/attestation-and-cryptographic-proofread 2026-09-25

The Investment and Wealth Workbench ships 23 named prebuilt agents for a customer to adopt, each built for one job and usable as shipped (8 research, 7 portfolio, 4 client communications, 4 compliance operations), on 95 workflows and 78 skills. Each stands on its own, and the Equity Screener, the Performance Attributor and the Risk and Scenario Analyst each do their own work without the others. The Workbench is in active design partner deployments and is included in the Prime tier.

SourceAlphaBitCore, alphabitcore.com (Workbench agents, home, pricing)read 2026-09-25

Four guardrail types, covering PII detection, content filtering, toxicity and banned words, can be chained into a pipeline for each agent. Three preconfigured risk profiles (low, medium and high) set an agent's starting trust tier. These are modules of one governance engine that the customer assembles, plus presets for settings. OpenBox names no catalog of ready made agents, workflows or policies.

SourceOpenBox, docs.openbox.ai/trust-lifecycle/authorize/guardrails and /trust-lifecycle/assessread 2026-09-25

Model choice and routing sit with the customer. The Nexus AI Gateway, the model layer of the AlphaBitCore stack, carries 20 provider codecs (OpenAI, Anthropic, Gemini, Vertex, Azure, Bedrock, Cohere and others, plus OpenAI compatible passthrough for Mistral, Groq, DeepSeek and more), routes across hosted and self hosted models, and policy checks every provider selection. The unified invocation contract makes switching providers a routing decision, with hosted and local models on one contract.

SourceAlphaBitCore, alphabitcore.com (nexus, platform/unify)read 2026-09-25

The agent's framework chooses its model, and OpenBox evaluates the operations that result. OpenBox offers no model choice of its own for the customer's agent. The OpenRouter SDK's routing policies are Rego rules that allow or deny model providers and record proof of where each call was routed. They govern the customer's model traffic and offer no choice inside OpenBox. The Alignment Score uses an LLM evaluation that OpenBox calls configurable, and it names no provider list or way to select one.

SourceOpenBox, docs.openbox.ai/developer-guide/openrouter/routing-policies and /core-concepts/trust-scoresread 2026-09-25

Outside callers can drive the platform. Prime MCP is the product's interface, with a published endpoint (https://finance.alphabitcore.io/mcp, Streamable HTTP, tenant format documented), a published auth scheme (OAuth 2.0 authorization code with PKCE and a documented discovery chain) and six enumerated tools that start real Prime tasks and answer approvals. Every model, agent, tool and workflow is also invoked through one SDK and HTTP contract, though the API reference is gated behind an evaluation.

SourceAlphaBitCore, alphabitcore.com (developers/mcp, developers, platform/unify)read 2026-09-25

SDKs for the platform include a Python Temporal plugin with a full API reference, Python SDKs for LangChain and LangGraph, CrewAI and Deep Agents, and TypeScript SDKs for Mastra and CopilotKit, plus an n8n community node and an OpenRouter SDK. Each comes with configuration options, error codes and an event model.

They authenticate to the hosted endpoint core.openbox.ai with API keys for each agent (live and test) and send each agent operation for a governance verdict. OpenBox names no REST API reference and no MCP server of its own. The MCP mentions in its docs belong to the Temporal demo agent.

SourceOpenBox, docs.openbox.ai/developer-guide/temporal-python/sdk-reference, /developer-guide/langchain and /getting-started/copilotkit/add-openbox-to-copilotkitread 2026-09-25

Nothing tests or scores what an agent produces. Deterministic replay and ValidatePoE verification prove that a recorded execution is complete and reruns to the same result, which is audit repeatability, not a test of output quality. Policies are code that can be diffed and tested, but there is no harness, scored test cases or quality gate for agent output, and the gateway benchmark is AlphaBitCore's own measurement of its product.

SourceAlphaBitCore, alphabitcore.com (platform/prove, platform/govern, nexus)read 2026-09-25

The Verify phase scores each session's actions against the agent's declared goal with an LLM evaluation that OpenBox calls configurable. Results show as an Alignment Score from 0 to 100 percent, with check counts, a trend line against a threshold (70 percent by default) and logged drift events.

The score tracks the customer's agent's behavior over time, and it runs only on live sessions. OpenBox names no way to test an agent against fixtures, mocks or datasets before production, and no red teaming before deployment or CI/CD checks. The Adapt phase tunes governance rules, not the agent, and the Assess phase is a risk questionnaire.

SourceOpenBox, docs.openbox.ai/trust-lifecycle/verify, /core-concepts/trust-scores and /trust-lifecycle/assessread 2026-09-25

No agent operates a browser or desktop. The closest thing is the Nexus gateway's Desktop Agent, which intercepts AI traffic at the operating system network layer (pf, iptables, WinDivert) so it passes through the compliance engine, but it navigates nothing.

SourceAlphaBitCore, alphabitcore.com (nexus)read 2026-09-25

Agent operations are governed, but OpenBox itself operates no browser or computer. Browsers appear only as steps in two demo walkthroughs.

SourceOpenBox, docs.openbox.ai/overview and the docs index (llms.txt and llms-full.txt)read 2026-09-25