[ PERSISTENT MEMORY INFRASTRUCTURE ]
Persistent shared memory for agentic systems. Hivemind recalls, compiles, and routes the right context before every turn.
EVERY MODEL STARTS FROM ZERO.
EVERY AGENT FORGETS WHAT THE LAST ONE LEARNED.
THAT IS NOT INTELLIGENCE.
THAT IS STATELESS AUTOMATION.
Agents write memories as immutable facts with rich metadata.
Hivemind searches the recall field for the most relevant context.
Context is ranked, deduped, and compiled into a task bundle.
The agent acts with full context and clear objectives.
Outcomes are recorded. Memories evolve. Hivemind gets smarter.
Checkout latency spiked during the payment provider failover; regional failback plus cache warmup restored p95 within 11 minutes.
Prefer reversible mitigations and preserve checkout stability during active traffic.
Artifacts: Checkout Latency Playbook v3.2; SLO dashboard shows p95 above threshold in us-east-1.
Last incident we deferred config changes and scaled the payment worker pool first.
Checkout change freeze: avoid changes that increase regression risk during active checkout traffic.
+ ranked, deduped, and trimmed before final assembly
Balanced admission inside max_context_tokens with per-class caps on memory, tool, and hold items.
Admitted items counted by source: hivemind, conversation, recalled_conversation, directive, system.
Every candidate gets an action: included, dropped_budget, dropped_policy, or trimmed_budget. Not prompt content.
Compiler diagnostics remain inspectable alongside the assembled bundle.
Context: 3220/4096 tokens | turn 14 | 13 items
Prefer reversible mitigations and preserve checkout stability during active traffic.
Find the lowest-risk path to stabilise checkout latency before peak traffic.
+ 10 more compiled messages: memory, artifacts, hold, and recalled conversation
{
"ok": true,
"data": {
"bundle": {
"messages": [
{ "role": "system", "content": "Context: 3220/4096 tokens | turn 14 | 13 items", "tool_call_id": null, "name": null },
{ "role": "system", "content": "=== AGENT FIXED CONTEXT BUFFER ===", "tool_call_id": null, "name": null },
{ "role": "system", "content": "Prefer reversible mitigations and preserve checkout stability during active traffic.", "tool_call_id": null, "name": null },
{ "role": "system", "content": "Artifacts: Checkout Latency Playbook v3.2; SLO dashboard shows checkout p95 above threshold in us-east-1.", "tool_call_id": null, "name": null },
{ "role": "system", "content": "Hold - Checkout change freeze: avoid changes that increase regression risk during active checkout traffic.", "tool_call_id": null, "name": null },
{ "role": "system", "content": "Checkout latency spiked during the 2026-06-21 payment provider failover; regional failback plus cache warmup restored p95 within 11 minutes.", "tool_call_id": null, "name": null },
{ "role": "system", "content": "Scaling the payment worker pool from 8 to 14 replicas resolved queue backpressure without config changes.", "tool_call_id": null, "name": null },
{ "role": "system", "content": "CDN config rollout on 2026-06-28 correlated with elevated checkout errors; rollback was clean and reversible.", "tool_call_id": null, "name": null },
{ "role": "system", "content": "Checkout Latency Playbook summary: prefer traffic shaping, replica scaling, and cache warmup before any config change.", "tool_call_id": null, "name": null },
{ "role": "assistant", "content": "Last incident we deferred config changes and scaled the payment worker pool first.", "tool_call_id": null, "name": null },
{ "role": "user", "content": "That worked - latency recovered inside the traffic window.", "tool_call_id": null, "name": null },
{ "role": "assistant", "content": "Monitoring is live. Ready for the next directive.", "tool_call_id": null, "name": null },
{ "role": "user", "content": "Find the lowest-risk path to stabilise checkout latency before peak traffic.", "tool_call_id": null, "name": null }
],
"items": [
{ "id": "context-status-14", "kind": "message", "source": "system", "role": "system", "timestamp": 1783261931.0, "priority": 950, "token_count": 36, "content": "Context: 3220/4096 tokens | turn 14 | 13 items", "metadata": { "turn": 14, "balance_lane": "fixed" } },
{ "id": "context-section-fixed-14", "kind": "message", "source": "system", "role": "system", "timestamp": 1783261931.0, "priority": 940, "token_count": 8, "content": "=== AGENT FIXED CONTEXT BUFFER ===", "metadata": { "section_label": "fixed", "section_title": "AGENT FIXED CONTEXT BUFFER" } },
{ "id": "context-briefing-14", "kind": "directive", "source": "directive", "role": "system", "timestamp": 1783261931.0, "priority": 1000, "token_count": 62, "content": "Prefer reversible mitigations and preserve checkout stability during active traffic.", "metadata": { "protected_core": true } },
{ "id": "context-artifacts-14", "kind": "message", "source": "system", "role": "system", "timestamp": 1783261931.0, "priority": 900, "token_count": 170, "content": "Artifacts: Checkout Latency Playbook v3.2; SLO dashboard shows checkout p95 above threshold in us-east-1.", "metadata": { "section_label": "artifacts" } },
{ "id": "hold-8c2f", "kind": "hold", "source": "system", "role": "system", "timestamp": 1783125600.0, "priority": 880, "token_count": 84, "content": "Checkout change freeze: avoid changes that increase regression risk during active checkout traffic.", "metadata": { "title": "Checkout change freeze", "tokens": 84, "held_since": 1783125600.0, "source": "operator", "source_id": "hold-8c2f" } },
{ "id": "memory-9d41c2", "kind": "memory", "source": "hivemind", "role": "system", "timestamp": 1782052800.0, "priority": 760, "token_count": 840, "content": "Checkout latency spiked during the 2026-06-21 payment provider failover; regional failback plus cache warmup restored p95 within 11 minutes.", "metadata": { "balance_lane": "memory", "utility_hint": 0.91 } },
{ "id": "memory-2ab7e4", "kind": "memory", "source": "hivemind", "role": "system", "timestamp": 1782398400.0, "priority": 740, "token_count": 720, "content": "Scaling the payment worker pool from 8 to 14 replicas resolved queue backpressure without config changes.", "metadata": { "balance_lane": "memory", "utility_hint": 0.87 } },
{ "id": "memory-77f0d3", "kind": "memory", "source": "hivemind", "role": "system", "timestamp": 1782657600.0, "priority": 720, "token_count": 610, "content": "CDN config rollout on 2026-06-28 correlated with elevated checkout errors; rollback was clean and reversible.", "metadata": { "balance_lane": "memory", "utility_hint": 0.83 } },
{ "id": "memory-4fb8aa", "kind": "memory", "source": "hivemind", "role": "system", "timestamp": 1782571200.0, "priority": 700, "token_count": 380, "content": "Checkout Latency Playbook summary: prefer traffic shaping, replica scaling, and cache warmup before any config change.", "metadata": { "balance_lane": "memory", "trimmed": true, "original_char_count": 5620, "trimmed_char_count": 1610 } },
{ "id": "recalled-conv-310", "kind": "message", "source": "recalled_conversation", "role": "assistant", "timestamp": 1783168200.0, "priority": 640, "token_count": 150, "content": "Last incident we deferred config changes and scaled the payment worker pool first.", "metadata": { "recalled": true } },
{ "id": "recalled-conv-311", "kind": "message", "source": "recalled_conversation", "role": "user", "timestamp": 1783168260.0, "priority": 630, "token_count": 120, "content": "That worked - latency recovered inside the traffic window.", "metadata": { "recalled": true } },
{ "id": "message-12", "kind": "message", "source": "conversation", "role": "assistant", "timestamp": 1783261870.0, "priority": 500, "token_count": 14, "content": "Monitoring is live. Ready for the next directive.", "metadata": {} },
{ "id": "message-13", "kind": "message", "source": "conversation", "role": "user", "timestamp": 1783261920.0, "priority": 500, "token_count": 26, "content": "Find the lowest-risk path to stabilise checkout latency before peak traffic.", "metadata": {} }
],
"decisions": [
{ "item_id": "context-briefing-14", "action": "included", "reason": null, "token_count": 62, "details": {} },
{ "item_id": "context-status-14", "action": "included", "reason": null, "token_count": 36, "details": {} },
{ "item_id": "context-section-fixed-14", "action": "included", "reason": null, "token_count": 8, "details": {} },
{ "item_id": "context-artifacts-14", "action": "included", "reason": null, "token_count": 170, "details": {} },
{ "item_id": "hold-8c2f", "action": "included", "reason": null, "token_count": 84, "details": {} },
{ "item_id": "memory-9d41c2", "action": "included", "reason": null, "token_count": 840, "details": {} },
{ "item_id": "memory-2ab7e4", "action": "included", "reason": null, "token_count": 720, "details": {} },
{ "item_id": "memory-77f0d3", "action": "included", "reason": null, "token_count": 610, "details": {} },
{ "item_id": "memory-4fb8aa", "action": "trimmed_budget", "reason": "token_budget", "token_count": 380, "details": { "original_char_count": 5620, "trimmed_char_count": 1610 } },
{ "item_id": "memory-c07d19", "action": "dropped_policy", "reason": "max_memory_items", "token_count": 590, "details": {} },
{ "item_id": "recalled-conv-309", "action": "dropped_budget", "reason": "token_budget", "token_count": 940, "details": {} },
{ "item_id": "recalled-conv-310", "action": "included", "reason": null, "token_count": 150, "details": {} },
{ "item_id": "recalled-conv-311", "action": "included", "reason": null, "token_count": 120, "details": {} },
{ "item_id": "message-12", "action": "included", "reason": null, "token_count": 14, "details": {} },
{ "item_id": "message-13", "action": "included", "reason": null, "token_count": 26, "details": {} }
],
"policy": {
"max_context_tokens": 4096,
"max_items": 50,
"max_memory_items": 4,
"max_tool_items": 10,
"max_hold_items": null,
"include_memory": true,
"include_tools": true,
"include_system": true,
"include_directives": true,
"include_holds": true,
"admission_strategy": "balanced"
},
"token_count": 3220,
"source_counts": { "hivemind": 4, "system": 4, "conversation": 2, "recalled_conversation": 2, "directive": 1 },
"budget_total": 4096,
"budget_available": 876,
"conversation_tokens": 310,
"hold_tokens": 84,
"trace": {
"session_id": "session_001",
"turn_number": 14,
"built_at": "2026-07-05T14:32:11+00:00",
"event_id": "event_001",
"trace_id": "trace_001",
"user_input": "Find the lowest-risk path to stabilise checkout latency before peak traffic.",
"budget_total": 4096,
"budget_used": 3220,
"budget_available": 876,
"conversation_tokens": 310,
"hold_tokens": 84,
"context_message_count": 13,
"source_counts": { "hivemind": 4, "system": 4, "conversation": 2, "recalled_conversation": 2, "directive": 1 },
"recalled_conversation_count": 2,
"recalled_memory_count": 4,
"recalled_memory_input_count": 7,
"semantic_memory_duplicates_filtered": 2,
"receipt_id": "e9b3c1a0-59f2-11f1-8d2e-0242ac120002",
"parent_receipt_id": null
}
},
"receipt": {
"receipt_id": "e9b3c1a0-59f2-11f1-8d2e-0242ac120002",
"operation": "context.compile",
"status": "ok",
"created_at": "2026-07-05T14:32:11+00:00",
"tenant_id": "tenant_acme",
"principal_id": "svc-runtime",
"agent_id": "agent_ops_01",
"session_id": "session_001",
"turn_id": "14",
"source_event_id": "event_001",
"trace_id": "trace_001",
"parent_receipt_id": null,
"request_id": "f21d7b6e-59f2-11f1-8d2e-0242ac120002",
"idempotency_key": null,
"inputs_ref": { "session_id": "session_001", "turn_number": 14, "budget_total": 4096, "recalled_conversation_count": 3, "recalled_memory_count": 7 },
"outputs_ref": { "token_count": 3220, "item_ids": ["context-briefing-14", "context-status-14", "context-artifacts-14", "hold-8c2f", "memory-9d41c2"], "source_counts": { "hivemind": 4, "system": 4, "conversation": 2, "recalled_conversation": 2, "directive": 1 } },
"metadata": { "trace_id": "trace_001", "event_id": "event_001" }
}
},
"request_id": "f21d7b6e-59f2-11f1-8d2e-0242ac120002"
}Not vector search. Purpose-built recall field that understands time, causality, and operator intent.
MORE SIGNAL. LESS NOISE.
One hive. All agents. Memories are agent-agnostic, deduped, and enriched as the hive learns.
NETWORK EFFECT FOR INTELLIGENCE.
Works with any model, framework, or runtime. Plug in via API. Keep your stack. Add a memory layer.
YOUR MODELS. OUR MEMORY.
Early access for teams building persistent agent systems.
Drop your email. We will be in touch when Hivemind launches.