ArkMemory

ArkMemory gives AI agents persistent memory across tasks using in-context, Redis, or vector store backends.

API: arkonis.dev/v1alpha1 Kind: ArkMemory Scope: Namespaced

By default, each agent task starts fresh — no knowledge of previous interactions. ArkMemory gives agents durable memory that persists across tasks and pod restarts. It declares where and how memory is stored, and ArkAgent claims it by name via spec.memoryRef.

How it works

The operator injects memory configuration into agent pods as environment variables. The agent runtime connects to the configured backend before polling the task queue.

During a task, the runtime:

Queries the memory backend for relevant past context and prepends it to the prompt
Writes the task result back to the memory backend after completion

The LLM never calls the memory backend directly — the runtime handles all reads and writes.

If the memory backend is unreachable at pod startup, the agent logs the error and continues without memory. Memory writes are best-effort and do not affect task delivery guarantees.

Backends

`in-context`

No external dependency. Memory is carried forward within a single task’s context window only. Once the task ends, the memory is gone. Use for stateless agents or prototyping.

`redis`

Short-term memory backed by Redis. Entries expire via a configurable TTL. Use for session-scoped context, recent interaction history, or deduplication state.

apiVersion: arkonis.dev/v1alpha1
kind: ArkMemory
metadata:
  name: research-memory
  namespace: default
spec:
  backend: redis
  redis:
    secretRef:
      name: redis-credentials   # must contain REDIS_URL
    ttlSeconds: 3600
    maxEntries: 500

`vector-store`

Long-term semantic memory backed by a vector database (Qdrant, Pinecone, or Weaviate). Past results are embedded and stored as vectors. At task start, the runtime performs a similarity search and retrieves the most relevant past context.

apiVersion: arkonis.dev/v1alpha1
kind: ArkMemory
metadata:
  name: research-memory-vector
  namespace: default
spec:
  backend: vector-store
  vectorStore:
    provider: qdrant
    endpoint: http://qdrant.agent-infra.svc.cluster.local:6333
    collection: agent-memories
    secretRef:
      name: qdrant-credentials
    ttlSeconds: 86400

Backend comparison

Backend	Best for	Notes
`in-context`	Stateless agents, prototyping	No external dependency.
`redis`	Short-term memory, session state	Fast reads and writes. Entries expire via TTL.
`vector-store`	Long-term semantic memory, RAG	Similarity search. Requires a running vector database.

Attaching memory to an agent

apiVersion: arkonis.dev/v1alpha1
kind: ArkMemory
metadata:
  name: research-memory
  namespace: default
spec:
  backend: redis
  redis:
    secretRef:
      name: redis-credentials
    ttlSeconds: 7200
---
apiVersion: arkonis.dev/v1alpha1
kind: ArkAgent
metadata:
  name: research-agent
  namespace: default
spec:
  model: llama3.2
  systemPrompt: "You are a research agent."
  memoryRef:
    name: research-memory

The same ArkMemory can be referenced by multiple ArkAgent resources in the same namespace.

Spec reference

`spec`

Field	Type	Required	Description
`backend`	string	yes	`in-context`, `redis`, or `vector-store`.
`redis`	RedisMemoryConfig	conditional	Required when `backend` is `redis`.
`vectorStore`	VectorStoreMemoryConfig	conditional	Required when `backend` is `vector-store`.

`spec.redis`

Field	Type	Default	Description
`secretRef.name`	string	required	Name of a Secret in the same namespace. Must contain a `REDIS_URL` key.
`ttlSeconds`	int	`3600`	How long memory entries are retained. `0` means no expiry.
`maxEntries`	int	`0` (unlimited)	Maximum entries per agent instance. Oldest entries are evicted when the limit is reached.

`spec.vectorStore`

Field	Type	Default	Description
`provider`	string	required	`qdrant`, `pinecone`, or `weaviate`.
`endpoint`	string	required	Base URL of the vector database.
`collection`	string	`agent-memories`	Collection or index name.
`secretRef.name`	string	no	Secret containing a `VECTOR_STORE_API_KEY` key. Required for hosted providers.
`ttlSeconds`	int	`0`	How long entries are retained. `0` means no expiry.

`status`

Field	Type	Description
`conditions`	[]Condition	`Ready=True` when the spec is valid and accepted.
`observedGeneration`	int64	The `.metadata.generation` this status reflects.

Environment variables injected

Variable	Description
`AGENT_MEMORY_BACKEND`	`in-context`, `redis`, or `vector-store`
`AGENT_MEMORY_REDIS_URL`	From the referenced Secret
`AGENT_MEMORY_REDIS_TTL`	TTL in seconds
`AGENT_MEMORY_REDIS_MAX_ENTRIES`	Entry cap
`AGENT_MEMORY_VECTOR_STORE_PROVIDER`	`qdrant`, `pinecone`, or `weaviate`
`AGENT_MEMORY_VECTOR_STORE_ENDPOINT`	Base URL
`AGENT_MEMORY_VECTOR_STORE_COLLECTION`	Collection name
`AGENT_MEMORY_VECTOR_STORE_API_KEY`	From the referenced Secret (if set)
`AGENT_MEMORY_VECTOR_STORE_TTL`	TTL in seconds

ArkMemory

How it works

Backends

in-context

redis

vector-store

Backend comparison

Attaching memory to an agent

Spec reference

spec

spec.redis

spec.vectorStore

status