| Core Platform | | | | |
| CTX Protocol + Parser ⓘ Seven operators × seven planes form a complete grammar for agent-to-system communication. Parser accepts CTX, JSON, or natural language. | | | | |
| 7 Context Planes ⓘ Tools, knowledge, memory, skills, agents, inspection, and LLM — covering every dimension of agent cognition. | | | | |
| CLI (actx) ⓘ Full CLI with 11+ subcommands: init, add, start, query, verify, top, ingest, watch, decode, phase, dispatch. | | | | |
| Ed25519 Signed Sidecar ⓘ Every sidecar translation is Ed25519-signed, creating a tamper-proof cryptographic receipt for audit verification. | | | | |
| MCP Server ⓘ Transparently proxies to any MCP-compatible backend. Agents interact with standard MCP while the gateway handles routing. | | | | |
| L1 Context Compaction ⓘ Active context window compaction — facts page in and out on demand instead of pre-loading everything. | | | | |
| Sidecar Emitters | | | | |
| Human (readable text) ⓘ Human-readable English translation of CTX operations. Enables non-technical stakeholders to audit agent behavior. | | | | |
| SurrealQL (DB queries) ⓘ Compiles CTX statements into SurrealQL queries for direct database execution with parameterized safety. | | | | |
| MCP-JSONRPC (tool dispatch) ⓘ Emits JSON-RPC calls for MCP tool invocations. Standard protocol for agent-to-tool communication. | | | | |
| CTXB (binary wire format) ⓘ Compact binary encoding achieving 1M+ ops/sec — ideal for high-throughput agent fleets. | | | | |
| REST (HTTP API) ⓘ Compiles CTX into RESTful HTTP calls for stateless integrations, webhooks, and serverless functions. | — | | | |
| GraphQL ⓘ Emits GraphQL queries and mutations for graph-based API integrations. | — | | | |
| SQL (generic) ⓘ Compiles CTX into standard SQL for compatibility with PostgreSQL, MySQL, and other relational databases. | — | | | |
| Reverse Parsing ⓘ Parses structured output back into CTX AST — enables round-trip translation and output verification. | — | | | |
| Infrastructure | | | | |
| Local SurrealDB Storage ⓘ SurrealDB serves as the unified backend for memory, knowledge, graph relations, and telemetry. | | | | |
| CTX Agent Memory (Cloud) ⓘ Cloud-hosted 5-layer memory architecture with perception, working, episodic, semantic, and procedural layers. | — | 10 GB | 50 GB | Dedicated |
| Native Surrealism Plugin ⓘ Same native Surrealism binary across all paid tiers — runs inside SurrealDB with zero JS bridge overhead. | — | | | |
| L2 Cross-Session CAS ⓘ Content-addressable storage with convergent encryption. Identical content deduplicates automatically across sessions. | — | | | |
| Remote MCP Backends ⓘ Connect to remote MCP servers via SSE transport. Enables cloud-hosted tools and cross-network agent capabilities. | — | | | |
| Hosted Gateway ⓘ Managed gateway with composable middleware: OIDC auth, RBAC, audit logging, alignment checks, and token budget tracking. | — | — | | |
| Dedicated Infrastructure ⓘ Dedicated SurrealDB and gateway instances — no noisy neighbors, sub-ms latency, full isolation. | — | — | — | |
| Security & Auth | | | | |
| Signed Payload Receipts ⓘ Ed25519 receipts on every sidecar operation. Creates an immutable audit trail connecting agent actions to translations. | | | | |
| Credential Vault ⓘ Resolves secrets from external vaults via op:// URI scheme with RAM-only caching and 15-minute TTL. | — | | | |
| Per-Tool RBAC ⓘ Every tool invocation is verified against the agent's RBAC role before dispatch. Prevents unauthorized tool access. | — | | | |
| Workspace RBAC ⓘ Role-based access control scoped to team workspaces. Isolates agent permissions across organizational boundaries. | — | — | | |
| Full RBAC + OIDC/SAML ⓘ Enterprise identity federation with OIDC auto-discovery, JWKS caching, and SAML integration. | — | — | — | |
| Sentinel Alignment ⓘ Pre/post execution hooks, behavioral heuristics, and cross-agent correlation with a graduated severity ladder. | — | — | — | |
| Audit Log Dashboard ⓘ Queryable audit trail with OpenTelemetry integration. Correlates agent actions across sessions and backends. | — | — | | |
| Cross-Tenant Federation ⓘ Identity-weighted consensus across organizational boundaries. Prevents Sybil attacks in federated deployments. | — | — | — | |
| Agents & Scale | | | | |
| Agent Limit ⓘ Maximum concurrent agents per account. Additional agents at $10/agent/month on Pro and Team tiers. | 1 | 5 | 25 | Unlimited |
| Session Persistence ⓘ Agent session state persistence. Cloud tiers store episodic and working memory in managed SurrealDB. | Local | Cloud | Cloud | Dedicated |
| Agent Identity ⓘ Verifiable agent identity for multi-agent coordination. Prevents Sybil attacks and enables trust-weighted decisions. | — | — | Workspace | Org-wide |
| Delegation ⓘ Agent-to-agent task delegation via structured handoff protocol. Transfers goals, progress, and working memory. | — | Single-user | Workspace | Cross-tenant |
| Shared Namespace ⓘ Team agents share a common memory and knowledge namespace. Enables collaborative context without manual sync. | — | — | | |
| Team Knowledge Sharing ⓘ Agents within a team can read and contribute to a shared knowledge plane. Collective intelligence across the org. | — | — | | |
| Fleet Dashboard ⓘ Real-time agent fleet monitoring — metrics, token budgets, cost reporting, and session telemetry in one view. | — | — | | |
| Orchestrator + Automation ⓘ Full lifecycle orchestration with event triggers, cron scheduling, webhooks, and checkpoint-based recovery. | — | — | — | |
| Support | | | | |
| Community Forum ⓘ Access to the AgentCTX community forum and GitHub discussions for peer support and knowledge sharing. | | | | |
| Email Support ⓘ Direct email support from the AgentCTX team with standard response times. | — | | | |
| Priority Support ⓘ Priority queue support with faster response times and escalation paths. | — | — | | |
| Dedicated Support + SLA ⓘ Dedicated support engineer, 99.9% uptime SLA, and guaranteed response times for mission-critical deployments. | — | — | — | |
| LLM Integration | | | | |
| Bring Your Own API Keys ⓘ Use your own LLM API keys — OpenAI, Anthropic, Google, or any provider. No markup on inference costs. | | | | |
| Managed LLM Routing ⓘ Automatic model selection based on task complexity and budget. Routes to the optimal provider for each operation. | — | Pay-per-use | Team rates | Volume pricing |
| Multi-Provider Failover ⓘ Automatic failover across LLM providers. If one provider is down or rate-limited, requests route to the next. | — | | | |
| Token Budget Management ⓘ Per-session and per-agent token budget enforcement with automatic alerts at configurable thresholds. | — | | | |
