Skip to content

Configuration Reference

This is the complete reference for all JarvisCore environment variables. Copy .env.example from your project root after running jarviscore init, populate one LLM section, and run jarviscore check to validate.


LLM Providers

Configure exactly one LLM provider. JarvisCore auto-detects the active provider from the environment variables present. The promotional provider is selected first when its token is present. The existing provider order after it is Azure → Claude → vLLM → Gemini → Vertex AI.

Existing Prescott promotional entitlement

Organizations that already received a Prescott promotional entitlement can set JARVISCORE_PROMO_TOKEN. This is a limited, revocable entitlement token, not an upstream model-provider API key. There is currently no public self-service enrollment page. Promotional access is selected first when configured. Expiry, exhaustion, and service errors fail explicitly and never silently fall through to another configured paid provider.

Every promotional call preserves the complete request and HTTP response under a stable call ID in ./.jarviscore/traces/promo_calls. Set JARVISCORE_PROMO_RAW_ARTIFACT_DIR to choose another local directory.

Variable Required Default Description
JARVISCORE_PROMO_TOKEN Yes (none) Unique promotional entitlement issued by Prescott
JARVISCORE_PROMO_RAW_ARTIFACT_DIR No ./.jarviscore/traces/promo_calls Durable complete call artifacts
Variable Required Default Description
CLAUDE_API_KEY Yes (none) API key starting with sk-ant-
CLAUDE_MODEL No claude-sonnet-4 Model name
Variable Required Default Description
GEMINI_API_KEY Yes (none) API key starting with AIza
GEMINI_MODEL No gemini-2.0-flash Model name
Variable Required Default Description
AZURE_API_KEY Yes (none) Azure OpenAI API key
AZURE_ENDPOINT Yes (none) Resource endpoint, e.g. https://your-resource.openai.azure.com/
AZURE_DEPLOYMENT Yes (none) Deployment name, e.g. gpt-4o
AZURE_API_VERSION No 2024-10-21 Azure OpenAI dated GA data-plane API version; override when your deployment requires another supported version
Variable Required Default Description
LLM_ENDPOINT Yes (none) OpenAI-compatible endpoint URL
LLM_MODEL Yes (none) Model name as expected by the endpoint

Multi-Tier Model Routing

The Kernel uses two base models that are always active, plus three optional tier overrides. Pass complexity= in workflow step dicts to select a tier.

await mesh.workflow("task-001", [
    {"agent": "analyst", "task": "...", "complexity": "nano"},
    {"agent": "analyst", "task": "...", "complexity": "heavy"},
])

Base models (always active)

Variable Default Description
CODING_MODEL dromos-gpt-4.1 Model used by CoderSubAgent for all code generation.
TASK_MODEL gpt-4o Model used by Researcher, Communicator, and Browser sub-agents when no tier override is set.

Tier overrides (optional)

Variable Tier Recommended use
TASK_MODEL_NANO nano Classification, summarisation, simple transforms
TASK_MODEL_STANDARD standard General tasks; the default when complexity is not specified
TASK_MODEL_HEAVY heavy Deep reasoning, long-context analysis, architecture decisions

When a tier variable is not set, the Kernel falls back to TASK_MODEL.

LLM reliability tuning

These become important once you run more than two agents concurrently against a rate-limited API.

Variable Default Description
LLM_TIMEOUT 120.0 Seconds before an LLM call times out.
LLM_CONNECT_TIMEOUT 10.0 Seconds to establish the connection (TCP and TLS), separate from LLM_TIMEOUT.
LLM_KEEPALIVE_SECONDS 120.0 Seconds an idle provider connection is kept for the next call (Azure, Claude, the promotion endpoint and vLLM). The SDK default of 5 s is shorter than the tool work between agent turns, so most calls reconnected; keep this below your provider's idle timeout (Azure: 240 s).
LLM_PREWARM true Open the primary provider's connection while an agent starts, so its first call skips the TCP and TLS handshake. Only a bare request to the provider's host is sent; a failed warm-up is ignored.
LLM_TEMPERATURE 0.7 Sampling temperature for all providers.
LLM_MAX_CONCURRENT 0 Most calls in flight per deployment. 0 leaves the governor to find the level (up to LLM_CONCURRENCY_CEILING). With the governor disabled, a process-wide cap.
LLM_MAX_RETRIES_429 4 Retry attempts when a provider returns 429 before giving up.
LLM_429_BASE_DELAY 2.0 Exponential backoff base delay in seconds. Actual delay: min(base × 2^attempt, 60s).
LLM_MODEL_OUTPUT_LIMITS {} JSON map of model or deployment name to its completion-token ceiling, e.g. {"gpt-5.2-chat": 128000}. Calls that do not set max_tokens receive the serving model's declared ceiling, bounded by the remaining workflow budget. Declare reasoning models here: their hidden reasoning is billed against the same allowance.
LLM_DEFAULT_MAX_TOKENS 4000 Output allowance for calls to models without a declared ceiling.
LLM_PRICES {} JSON map of model or deployment name to its price per 1M tokens, e.g. {"gpt-5.6-luna": {"input": 1.25, "output": 10.0, "cached": 0.125}}. Declared prices override the built-in table. Cached input is billed at cached when given. A model with no known price is costed at $3/$15 per 1M input/output tokens, logged once as an estimate, and its calls report price_source: "default".
AZURE_CONTENT_FILTER_REPAIR_ENABLED false Opt into an Azure-specific content-filter retry that applies a provider-safe preamble and neutral wording after the raw prompt is rejected. Off by default so prompt rewriting never hides developer intent.

Provider governor

Every model call is admitted per deployment (for example azure:gpt-4o) before it is sent. Calls in flight start at LLM_INITIAL_CONCURRENCY, grow while the provider keeps up, and halve on a 429 or near-empty rate-limit headers. None of these needs setting.

Variable Default Description
LLM_GOVERNOR_ENABLED true Admit every call through the governor.
LLM_LIMITS {} Known limits per deployment, enforced in a 60 s window shared through Redis by every process, e.g. {"azure:gpt-4o": {"rpm": 3000, "tpm": 1000000}}.
LLM_INITIAL_CONCURRENCY 8 Calls in flight per deployment at start.
LLM_CONCURRENCY_CEILING 256 Most calls in flight per deployment the governor will grow to.
LLM_BULK_RESERVE 0.15 Share of a deployment's calls in flight held back from bulk work for interactive calls.

Run background work in the bulk lane so people waiting on chat keep headroom:

from jarviscore.execution.governor import BULK, lane_scope

with lane_scope(BULK):
    await mesh.execute_goal("Re-verify every claim in the data room")

Decision Models: TypeSafe Jev

TypeSafe Jev is an optional decision model for typed Choice, Score, and Noul judgments. Install jarviscore-framework[typesafe]. AutoAgent still requires one generative LLM provider from the preceding section.

Variable Default Description
DECISION_CLIENT_FACTORY (none) module:callable returning any DecisionClient; takes precedence over TYPESAFE_API_KEY. See Bring any decision model.
TYPESAFE_API_KEY (none) Enables the shared Jev decision client.
TYPESAFE_DEFAULT_MODEL jev-latest TypeSafe model used for decisions.
TYPESAFE_BASE_URL TypeSafe SDK default Optional compatible gateway or test endpoint.
TYPESAFE_TIMEOUT 10.0 Decision request timeout in seconds.
KERNEL_ROUTER_PROVIDER llm Set to typesafe to opt into Jev-backed Kernel role selection.
KERNEL_ROUTER_MAY_ANSWER true Let the router answer, in the same call, a task that needs no role (summarising, rewording, explaining or drafting from the task text, with no tool, data, calculation or action). Applies only when the router runs on the model the communicator would use.
TYPESAFE_ROUTER_MIN_CONFIDENCE 0.5 Minimum Jev routing confidence. Validate against your workload before changing production behavior.
TASK_COMPLEXITY_PROVIDER llm Set to typesafe to classify direct Kernel versus Planner execution with Jev.
TYPESAFE_COMPLEXITY_MIN_CONFIDENCE 0.5 Below this confidence, preserve the full planning path.
RAG_DECISION_PROVIDER vector Set to typesafe to classify the FAISS shortlist before Researcher consumes it.
RAG_TYPESAFE_MAX_CONCURRENT 4 Maximum concurrent passage-classification calls.
RAG_TYPESAFE_INJECTION_MAX 0.70 Exclude passages above this prompt-injection probability. Advisory, not a security boundary.
RAG_TYPESAFE_CONTRADICTS_MIN 0.70 Route passages above this premise-conflict probability to the conflict set.
RAG_TYPESAFE_RELEVANT_MIN 0.45 Exclude passages below this relevance probability.
RAG_TYPESAFE_EVIDENCE_MIN 0.55 Include passages above this usable-evidence probability after earlier checks.
TYPESAFE_INPUT_COST_PER_MILLION_USD 0.042 Input-token price used for workflow cost accounting; update if TypeSafe pricing changes.

See Decision Models for agent APIs, routing precedence, privacy boundaries, and examples.


Redis

Without Redis, JarvisCore runs in in-process mode: workflows execute locally, mailboxes are file-backed, and distributed features are unavailable. With Redis, the following capabilities become active: distributed workflow DAGs, durable mailboxes, cross-node step claiming, episodic ledger, and LTM compression.

Variable Default Description
REDIS_URL (none) Full connection URL; takes precedence over host/port variables. Example: redis://host:6379/0
REDIS_HOST localhost Used when REDIS_URL is not set
REDIS_PORT 6379 Used when REDIS_URL is not set
REDIS_PASSWORD (none) Authentication password
REDIS_DB 0 Database number
REDIS_CONTEXT_TTL_DAYS 7 Number of days to retain agent context keys

Start Redis locally for development:

docker run -d -p 6379:6379 redis:7-alpine

Distributed goal configuration

These values are Mesh(config={...}) keys, not environment variables:

Key Default Description
distributed_poll_interval 2.0 Redis DAG polling interval in seconds
distributed_claim_lease_seconds 60 Renewable execution-claim lease duration
mesh_planning_lease_seconds 300 Initial planning and amendment lease duration
mesh_max_reconciliation_revisions 3 Bounded automatic semantic revisions
mesh_response_capability None Capability authorized for the optional final response
mesh_planning_brief None Trusted product mandate, operating method and negative-result evidence standard used for DAG draft, audit, repair and reconciliation; never used to extract source obligations
execution_budget framework defaults Shared max_seconds, max_steps, max_replans, max_tokens, max_epochs_per_step, max_peer_depth and peer_timeout_seconds
workspace_required false Fail before planning unless a trusted source adapter produces a verified snapshot
workspace_source_adapters {} Provider names mapped to SourceAdapter implementations
workspace_source_catalog built-in provider descriptions Trusted provider descriptions used to resolve explicit source identity from a goal
workspace_source_resolver planning-model resolver Optional trusted callable that returns source identity before adapter capture
workspace_snapshot_limits framework defaults Trusted max_files, max_total_bytes, max_file_bytes, and acquisition concurrency limits enforced before planning
workspace_allowed_commands [] Additional trusted executable names available to workspace_run; caller context cannot expand this list. A node that sets it decides for every step it runs, whoever asked
workspace_command_timeout_seconds 120 Trusted per-command workspace limit; a timeout is evidence and does not stop the DAG
workspace_command_environment {} Trusted recognized package-manager cache paths passed without inheriting parent secrets
workspace_delta_ignored_names runtime/build defaults Directory names excluded from persisted deltas, including .git, .venv, node_modules, output, and target

Caller task context cannot override execution authority or the workflow budget. All nodes sharing one distributed DAG should run the same JarvisCore minor version. See Durable Goal Execution.


Memory: Athena MemOS

Athena provides three-tier persistent memory (STM, MTM, and LTM graph) that spans sessions. Without Athena, agents use Redis-only episodic memory for the current session.

Variable Default Description
ATHENA_URL (none) Athena API base URL, e.g. http://localhost:8080
ATHENA_TENANT_ID default Tenant namespace for memory isolation across teams or environments
ATHENA_HTTP_TIMEOUT 10.0 Seconds before an Athena HTTP call times out. Increase if your Athena instance is remote or under load.
ATHENA_SESSION_TTL_DAYS 30 How long a session ID is cached in Redis before Athena re-issues it.
ATHENA_API_KEY — Optional API key sent as X-API-Key. Keep it out of prompts and logs.
ATHENA_JWT_TOKEN — Optional JWT sent as X-JWT-Token for tenant/user isolation.
ATHENA_PORT 8080 Host port for the Athena server when the stack is started by memory init

Every published port in the local stack is overridable for co-located deployments: ATHENA_REDIS_PORT (6380), ATHENA_MONGO_PORT (27017), ATHENA_MINIO_PORT (9000), ATHENA_MINIO_CONSOLE_PORT (9001), ATHENA_MILVUS_PORT (19530), ATHENA_MILVUS_METRICS_PORT (9091), ATHENA_ARANGO_PORT (8529).

Azure OpenAI is detected automatically by memory init when AZURE_OPENAI_API_KEY and AZURE_OPENAI_ENDPOINT are set (deployment name from AZURE_DEPLOYMENT, default gpt-4o).

Initial setup:

git clone https://github.com/Prescott-Data/athena ~/athena
jarviscore memory init

If your Athena clone is not at ~/athena:

ATHENA_DIR=/path/to/athena jarviscore memory init

Storage: Blob

Blob storage is used by agents to persist large outputs (reports, datasets, generated files) and by WorkingScratchpad for per-step working notes. The local backend is always active by default.

Variable Default Description
STORAGE_BACKEND local local, azure, or redis (file contents in REDIS_URL, so one Redis serves every machine)
STORAGE_BASE_PATH ./.jarviscore/blob_storage Base directory path for the local backend
STORAGE_REDIS_MAX_BLOB_BYTES 33554432 Largest single file the redis backend accepts
AZURE_STORAGE_CONNECTION_STRING (none) Required when STORAGE_BACKEND=azure

For Azure:

pip install "jarviscore-framework[azure]"
.env
STORAGE_BACKEND=azure
AZURE_STORAGE_CONNECTION_STRING=DefaultEndpointsProtocol=https;AccountName=...

Provider deadlines

Variable Default Description
RESEARCH_GROUNDED_TIMEOUT_SECONDS 45 Per-call deadline for Google Grounded (generative search)
RESEARCH_SEARCH_TIMEOUT_SECONDS 15 Per-call deadline for each other search provider
RESEARCH_ALLOW_WIKIPEDIA_FALLBACK false Search Wikipedia when no other provider found anything, even with web providers configured

Both InternetSearch implementations accept grounded_timeout_seconds and search_timeout_seconds keyword arguments, which take precedence over these environment variables. Values must be positive finite seconds. For example, InternetSearch(grounded_timeout_seconds=60, search_timeout_seconds=20). These deadlines bound each provider including its retries; existing HTTP-request timeouts still apply. They do not change agent leases or the overall task budget. Timeout diagnostics identify the provider, exception type and configured deadline.

Crawler identity

Variable Default Description
RESEARCH_USER_AGENT A desktop browser string User-Agent sent when research fetches pages and PDFs. Public registers such as SEC EDGAR admit automated clients only when they declare who they are, for example Acme Research research@acme.example. Both InternetSearch implementations also accept a user_agent argument.

JarvisCore runs multiple search providers in parallel and merges results. All providers have circuit breakers: a failing provider is skipped automatically. See the Internet Search guide for provider details, ranking logic, and usage patterns.

Google Grounded Search Primary

Active automatically when GEMINI_API_KEY is set (which is already required for agents). No extra configuration needed.

Variable Default Description
GEMINI_GROUNDING_API_KEY GEMINI_API_KEY Override the key used specifically for grounded search
GEMINI_GROUNDING_MODEL gemini-2.5-flash Gemini model used for grounded search
GOOGLE_CLOUD_PROJECT (none) Vertex AI path (alternative to API key)
GOOGLE_CLOUD_LOCATION global GCP location for Vertex AI

Serper Optional

Variable Default Description
SERPER_API_KEY (none) API key from serper.dev. Provider is skipped if unset.

SearXNG Optional Self-hosted

Variable Default Description
SEARXNG_INSTANCE_URL http://localhost:8080 URL of your SearXNG instance. Provider is skipped if unreachable.

Research & PDF tuning

Variable Default Description
RESEARCH_PDF_TIMEOUT_SECONDS 90 Timeout for PDF download and extraction
RESEARCH_PDF_MAX_RETRIES 3 Retry attempts for PDF extraction

Browser Automation

Enables web navigation, form interaction, and structured scraping via BrowserSubAgent.

pip install "jarviscore-framework[browser]"
playwright install chromium
Variable Default Description
BROWSER_ENABLED false Enable BrowserSubAgent
BROWSER_HEADLESS true Run the browser in headless mode
BROWSER_SITE_INTERVAL_S 1.5 Seconds between requests to the same site
BROWSER_SETTLE_TIMEOUT_MS 3000 How long a page may keep changing after an action before the agent looks again
BROWSER_DATA_TIMEOUT_MS 10000 Longest an action waits for the site's own data requests to finish
BROWSER_CODE_TIMEOUT_S 600 Default and maximum program duration; also bounded by the remaining execution lease
BROWSER_PARALLEL_PAGES 4 Pages read_pages opens at once in a run's browser session
BROWSER_HISTORY_PAGES 50 Most pages read_pages turns through one history before reporting that it goes on (more)
BROWSER_ACT_MIN_CONFIDENCE 0.7 Decision confidence at which act performs the action itself; below it, act returns candidates
BROWSER_CONTINUATION_HOLD_S 120 How long a browser stays open for the step's next execution epoch to continue on the same page
BROWSER_LOCALE host's Locale the browser presents, e.g. en-GB; match it to the egress proxy's region
BROWSER_TIMEZONE host's IANA timezone the browser presents, e.g. Europe/London

Egress Proxy

JarvisCore supplies no IP addresses. To give agents a reputable, stable or regional exit, point them at your own proxy provider. The browser and the pages the researcher reads leave through it; search-provider APIs, LLM calls and Nexus do not.

Variable Default Description
EGRESS_PROXY_URL (none) http:// or https:// proxy, e.g. http://user-session-{session}:pass@gw.provider.example:7000. {session} becomes a stable 16-character id per tenant profile or workflow, so a provider that pins an exit IP per session keeps one IP for a login or checkout.
EGRESS_PROXY_BYPASS (none) Comma-separated hosts that go direct, e.g. searxng,.internal. Loopback always goes direct.

See Browser Automation: Your own proxy.


Nexus: Credential Management

Required only when agents call third-party services (GitHub, Slack, Jira, Stripe, and so on) through providers registered with requires_auth = True.

Variable Default Description
NEXUS_GATEWAY_URL (none) Gateway URL, e.g. http://localhost:8090
NEXUS_GATEWAY_API_KEY (none) Sent as X-API-Key on every Gateway call. Written by jarviscore nexus init. Grants Gateway access only, never the broker's encryption key.
NEXUS_BROKER_URL (none) Broker URL, e.g. http://localhost:8080. Used by dashboard and SDK to POST credentials directly to the Broker's /auth/capture-credential endpoint for non-OAuth providers.
NEXUS_RETURN_URL http://localhost:8000/oauth/callback OAuth callback URL; the Nexus broker redirects here after consent.
NEXUS_DEFAULT_USER_ID jarviscore-agent User identity passed to Nexus for credential lookups.
NEXUS_SUBJECT_BINDING optional For jarviscore nexus up: the broker's SUBJECT_BINDING. required releases a credential only to work naming the person who owns it; use it when agents act for your users (guide).
AUTH_STRATEGY_CACHE_TTL 300 Seconds before the resolved auth strategy is re-fetched from the Gateway.
AUTH_FLOW_TIMEOUT 300 Maximum seconds to wait for a user to complete OAuth browser consent.
AUTH_POLL_INTERVAL 2.0 Seconds between Gateway status polls during an OAuth flow.
AUTH_OPEN_BROWSER true Whether to open the system browser automatically when OAuth consent is needed.

Initial setup:

jarviscore nexus init
jarviscore nexus register github --client-id=YOUR_ID   # prompts for the secret
jarviscore nexus test github

The local stack's own secrets (NEXUS_ENCRYPTION_KEY, NEXUS_STATE_KEY, NEXUS_BROKER_API_KEY) live in ~/.jarviscore/nexus/stack.env and are read by the broker and gateway containers only. Do not put them in a project .env.

Control and evidence

How provider calls are decided on and recorded. See Control and Evidence.

Variable Default Description
CONTROL_MODE report off: provider calls and other effects are neither decided on nor recorded. report: decided and recorded; refused only when beyond what the person the work is for may do (mandate, scopes). strict: refused before they happen when policy does not allow them, and refused when their decision cannot be recorded; needs a confined sandbox for model-written code. JARVISCORE_CONTROL_MODE overrides it at run time. See Turning On Control.
CONTROL_POLICY_PATH (none) JSON or YAML file of deployment rules: deny rules refuse more; an allow rule can lift only the refusal of a mutation that does not come from an integration. approvals: decides who may approve; grounding: holds request fields to records the runtime read; effects: decides on every other effect; principals: gives approvers roles; scopes: says which operations need which of your organisation's scopes. Full schema: Control Policy Reference; check a file with jarviscore policy check, list who may do what with jarviscore policy scopes. A file that cannot be read refuses every call and every effect. JARVISCORE_CONTROL_POLICY overrides it at run time.
JOURNAL_DIR .jarviscore/journal Where the hash-linked journal of decisions, provider calls and other effects is written; off writes none (and strict mode then cannot record its decisions). Check it with jarviscore verify; jarviscore policy review lists what strict mode would refuse; jarviscore evidence <person> what was done for someone. JARVISCORE_JOURNAL_DIR overrides it at run time.
JOURNAL_CAPTURE full What each agent run keeps for replay: full (model calls, gate and tool results, content-addressed beside the journal), digests (verifiable, not replayable) or off. JARVISCORE_JOURNAL_CAPTURE overrides it. See Replay.
JOURNAL_CONTENT_DAYS 30 Days replay content is kept after its last use; older content is erased in the background, at most once a day, by processes that record agent runs with full capture. 0 keeps it. The journal still verifies; those runs can no longer be replayed. JARVISCORE_JOURNAL_CONTENT_DAYS overrides it.
NEXUS_KEEPALIVE_SECONDS 120 Seconds an idle provider connection is kept for the next call, so calls reuse TLS instead of opening a connection each time.
APPROVAL_ASSERTION_KEY (none) Key approval assertions are verified with (key id default). Your approving backend holds the same key and signs with control.approvals.sign_assertion after its own sign-in.
APPROVAL_ASSERTION_KEYS {} JSON map of key id to key, for rotating assertion keys.

P2P / SWIM Mesh

Enables multi-node agent discovery and message routing using the SWIM gossip protocol and ZMQ transport. Requires Redis for distributed workflow coordination.

Variable Default Description
P2P_ENABLED true Enables SWIM/ZMQ when the optional p2p dependencies are installed; set false to force local-only peers
JARVISCORE_BIND_HOST 127.0.0.1 Per-process bind address; use 0.0.0.0 for a node reachable from other machines
JARVISCORE_BIND_PORT 7946 Per-process SWIM gossip port; ZMQ uses this port plus 1000
JARVISCORE_SEED_NODES (none) Comma-separated seed addresses, for example 10.0.0.1:7946,10.0.0.2:7946

Port uniqueness

JARVISCORE_BIND_PORT must be different for each node running on the same machine. The ZMQ data port is set automatically to JARVISCORE_BIND_PORT + 1000.

The seed node does not set JARVISCORE_SEED_NODES. All other nodes point at the seed node (or any other live node) to join the cluster. Keep these bind settings per process rather than in a shared .env file.


Kernel Tuning

Advanced execution parameters for the OODA loop Kernel. The defaults are appropriate for the majority of workloads.

Execution limits

Variable Default Description
KERNEL_MAX_TURNS 30 Maximum OODA loop turns per task execution.
SANDBOX_MODE local local (in-process execution) or remote (external sandboxed execution).
ALLOW_UNSAFE_LOCAL_EXECUTION false Run model-written code as an unconfined host subprocess where OS confinement is unavailable. With the default, such code runs only under macOS sandbox-exec or Linux bwrap (no network, writes only to the run workspace) and is refused elsewhere. Keep disabled for untrusted agent input; see Sandbox Execution.
EXECUTION_TIMEOUT 300 Seconds before sandbox code execution is forcibly terminated.
MAX_REPAIR_ATTEMPTS 3 Autonomous repair retries when generated code fails.
HITL_ENABLED false Enable Human-in-the-Loop escalation.
HITL_MAX_CONFIDENCE 0.8 Escalate when the Kernel's action confidence is below this value.
HITL_MIN_RISK_SCORE 0.7 Escalate when the evaluated risk score exceeds this value.
HITL_TARGETS reviewer Who responds to escalations, comma-separated (e.g. ops-oncall,finance-approver). Carried on every request as targets; a request can name its own. See Who responds.
MAX_GOAL_STEPS 30 Hard ceiling on plan steps for goal-oriented agents before the goal is marked blocked.
MAX_REPLAN_ATTEMPTS 8 Maximum replanning cycles before the Kernel returns a failure result.

Token budgets

These control how the Kernel allocates context across its OODA turns. You will not need to change these unless you are working with very long tasks or seeing silent truncation.

Variable Default Description
KERNEL_MAX_TOTAL_TOKENS 80000 Hard token ceiling per task execution across all Kernel turns.
KERNEL_THINKING_BUDGET 56000 Token allocation for Kernel reasoning turns.
KERNEL_ACTION_BUDGET 24000 Token allocation for action execution turns.
KERNEL_WALL_CLOCK_MS 180000 Maximum wall-clock time in milliseconds for a single task execution (3 minutes).

Tool result observation

Each tool result is shown to the agent whole, once: in the conversation while it is among the agent's recent exchanges, and in the list of recent actions after that. Results of the last few turns also stay readable with read_turn_result, a page at a time. Nothing is cut: older results are released whole, and the agent is told why.

Variable Default Description
SUBAGENT_OBSERVATION_LIMIT removed Tool results are no longer clipped; the agent sees each one whole. The setting is ignored.
SUBAGENT_TURN_RESULT_WINDOW 10 Number of recent tool results kept whole for read_turn_result.
SUBAGENT_TURN_RESULT_BUDGET 2000000 Total characters kept across those results; older results are released whole when exceeded.
SUBAGENT_READ_PAGE_LIMIT 20000 Largest page read_turn_result returns; each page is shown whole.

Agent Profiles

Variable Default Description
JARVISCORE_PROFILES_DIR Bundled fallback Absolute path to your application's agent profile YAML directory

Set this to your application's profiles directory so that agents load domain-specific role intelligence rather than the bundled example profiles:

.env
JARVISCORE_PROFILES_DIR=/path/to/your-app/profiles/agents

Mailbox

The MailboxManager is the Redis Streams-backed fire-and-forget messaging layer between agents. These settings only apply when Redis is configured.

Variable Default Description
MAILBOX_MAX_MESSAGES 100 Maximum messages drained per read() call.
MAILBOX_POLL_INTERVAL 0.5 Seconds between mailbox poll cycles in the listener loop.

Function Registry

The FunctionRegistry (atom store) promotes generated code through three quality levels: Candidate, Verified, and Golden. These thresholds control when promotion occurs.

Variable Default Description
REGISTRY_VERIFIED_THRESHOLD 1 Successful executions before a Candidate atom is promoted to Verified.
REGISTRY_GOLDEN_THRESHOLD 5 Successful executions before a Verified atom is promoted to Golden. Golden atoms are always tried first.
REGISTRY_MAX_CACHE_SIZE 500 Maximum atoms held in the in-memory registry cache.

See System Bundles and Atoms for a full explanation of the graduation model.


Telemetry

Variable Default Description
TELEMETRY_ENABLED true Write execution trace files to disk.
TELEMETRY_TRACE_DIR ./.jarviscore/traces Directory for execution trace files.
PROMETHEUS_ENABLED false Start a Prometheus metrics endpoint.
PROMETHEUS_PORT 9090 Port for the /metrics endpoint.
pip install "jarviscore-framework[prometheus]"

Logging

Variable Default Description
LOG_LEVEL INFO Log verbosity level: DEBUG, INFO, WARNING, or ERROR