# Akhara

> Akhara AI verification suite: Environments (EaaS), Control plane (runtime policy), Evaluation (domain-agnostic AI evals + CI/CD), and Expert Review (audit/review with BYO experts; docs path `talent/`).

## Agent instructions

- Company name is **Akhara AI** (never Rubric AI). Keep lowercase **rubric/rubrics** only when meaning grading criteria.
- Prefer concrete API examples against the public hosts listed below. Do not invent a public hostname for private orchestrators.
- Environments: Eval API `https://agi.akhara.ai`, Admin `https://admin-agi.akhara.ai`. Use env ABI terms (environment, eval API, orchestrator); label legacy fields like `harbor_reset` as legacy.
- Control plane: PDP `https://api.akhara.dev`, Console `https://console.akhara.dev`. Do not confuse control-plane latches with Environments confirmation latches.
- Evaluation: App `https://app.akhara.ai`. SDK uses `from akhara import Akhara` and `AKHARA_API_KEY`.
- Expert Review (path `talent/`): Portal `https://talent.akhara.ai`: enterprises invite their own specialists to audit runs, review hard cases, and improve tasks and rubrics; RLHF evaluations and domain writing for RL environments are secondary work types.
- Fetch individual pages as Markdown by appending `.md` to the doc path (Mintlify export).
- For the full concatenated export when available, see `/llms-full.txt`.

## Start here

- [Akhara AI](/index.md): Akhara AI - verification layer for AI agents: why Akhara, the verification stack, and when to use each product
- [Glossary](/home/glossary.md): Cross-product terms (control plane, Environments, Evaluation, Expert Review)
- [Tenancy and security](/security/tenancy.md): Isolation bar, quotas, and defaults across products

## Environments

- [Akhara Environments](/environments/index.md)
- [Quickstart](/environments/quickstart.md)
- [Architecture](/environments/architecture.md)
- [API overview](/environments/api/overview.md)
- [Evaluations API](/environments/api/evaluations.md)
- [Env ABI](/environments/api/env-abi.md)
- [Dashboard API](/environments/api/dashboard.md)
- [Verifiers](/environments/mechanics/verifiers.md)
- [Confirmation latches](/environments/mechanics/confirmation-latches.md)
- [Scores and stats](/environments/results/scores-and-stats.md)
- [Tenancy & security](/security/tenancy.md)
- [Environment](/environments/mechanics/environment.md)
- [Task definitions](/environments/mechanics/task-definitions.md)
- [Actions](/environments/mechanics/actions.md)
- [Sample trajectories](/environments/results/sample-trajectories.md)

## Control plane

- [Control plane](/control-plane/index.md)
- [Quickstart](/control-plane/quickstart.md)
- [Overview](/control-plane/concepts/overview.md)
- [Latching](/control-plane/concepts/latching.md)
- [API reference](/control-plane/api-reference/introduction.md)
- [Authorize a step](/control-plane/api-reference/endpoint/authorize.md)
- [SDK requirements](/control-plane/sdk/requirements.md)
- [Tenancy & security](/security/tenancy.md)
- [Architecture](/control-plane/concepts/architecture.md)
- [Verdicts](/control-plane/concepts/verdicts.md)
- [Fail-closed](/control-plane/concepts/fail-closed.md)
- [Agent onboarding & connection](/control-plane/onboarding/agents.md)
- [Policies](/control-plane/onboarding/policies.md)
- [Verifier crafting](/control-plane/onboarding/verifiers.md)
- [Task publishing](/control-plane/onboarding/tasks.md)
- [Environment provisioning](/control-plane/onboarding/environments.md)
- [Running](/control-plane/onboarding/running.md)
- [TypeScript SDK](/control-plane/sdk/typescript.md)
- [Kotlin / Android SDK](/control-plane/sdk/kotlin.md)
- [List agents](/control-plane/api-reference/endpoint/list-agents.md)
- [Create / connect an agent](/control-plane/api-reference/endpoint/create-agent.md)
- [Attach policies](/control-plane/api-reference/endpoint/set-policies.md)
- [Provision a runtime](/control-plane/api-reference/endpoint/provision-agent.md)
- [Delete an agent](/control-plane/api-reference/endpoint/delete-agent.md)
- [Ingest a telemetry event](/control-plane/api-reference/endpoint/ingest-event.md)
- [Seal a decision](/control-plane/api-reference/endpoint/sign-record.md)
- [Verify a sealed decision](/control-plane/api-reference/endpoint/verify-record.md)

## Evaluation

- [Introduction to Akhara Evaluation](/evaluation/docs/introduction.md)
- [Quickstart](/evaluation/docs/getting-started/quickstart.md)
- [Deploy evaluations in CI/CD](/evaluation/docs/tutorials/ci-cd.md)
- [API Reference](/evaluation/api-reference/introduction.md)
- [Authentication Overview](/evaluation/api-reference/authentication/overview.md)
- [Datasets Overview](/evaluation/api-reference/datasets/overview.md)
- [Evaluations Overview](/evaluation/api-reference/evaluations/overview.md)
- [Rubrics Overview](/evaluation/api-reference/rubrics/overview.md)
- [Python SDK](/evaluation/api-reference/sdks/python.md)
- [Platform capabilities](/evaluation/docs/platform-capabilities.md)
- [Open-source tooling](/evaluation/docs/open-source.md)
- [Akhara vs. alternatives](/evaluation/docs/comparison.md)
- [Account Setup](/evaluation/docs/getting-started/setup.md)
- [Create Your First Evaluation](/evaluation/docs/getting-started/first-evaluation.md)
- [Evaluation Quickstart](/evaluation/docs/getting-started/evaluation.md)
- [Observability Quickstart](/evaluation/docs/getting-started/observability.md)
- [Core Objects](/evaluation/docs/core-concepts/objects.md)
- [Evaluation Lifecycle](/evaluation/docs/core-concepts/evaluation-lifecycle.md)
- [Human vs Automated Evaluation](/evaluation/docs/core-concepts/human-vs-automated.md)
- [System Architecture Overview](/evaluation/docs/architecture/overview.md)
- [Data Flow Diagram](/evaluation/docs/architecture/data-flow.md)
- [Human Expert Routing Engine](/evaluation/docs/architecture/routing-engine.md)
- [PHI/PII Handling Model](/evaluation/docs/architecture/phi-pii-handling.md)
- [Audit & Provenance Pipeline](/evaluation/docs/architecture/audit-provenance.md)
- [Evaluating LLM Applications](/evaluation/docs/onboarding/evaluating-llm.md)
- [Evaluating Voice & Multimodal AI](/evaluation/docs/onboarding/voice-multimodal.md)
- [Evaluation framework](/evaluation/docs/evaluation-framework/overview.md)
- [Evaluation Types](/evaluation/docs/evaluation-framework/types.md)
- [Metrics Reference](/evaluation/docs/evaluation-framework/metrics.md)
- [Evaluation Versioning](/evaluation/docs/evaluation-framework/versioning.md)
- [Human Review Design](/evaluation/docs/evaluation-framework/human-review-design.md)
- [Human Expert Network](/evaluation/docs/evaluation-framework/human-expert-network.md)
- [Evaluate a Medical Voice Agent](/evaluation/docs/tutorials/voice-agent.md)
- [Evaluate RAG for Clinical Summaries](/evaluation/docs/tutorials/rag-evaluation.md)
- [Safety Gating Before Production](/evaluation/docs/tutorials/safety-gating.md)
- [Continuous Monitoring in Production](/evaluation/docs/tutorials/monitoring.md)
- [Export Results for Regulators](/evaluation/docs/tutorials/export.md)
- [Integrations Overview](/evaluation/docs/integrations/overview.md)
- [LLM Providers](/evaluation/docs/integrations/llm-providers.md)
- [EHR Integration](/evaluation/docs/integrations/ehr.md)
- [FHIR Integration](/evaluation/docs/integrations/fhir.md)
- [Data Warehouses](/evaluation/docs/integrations/data-warehouses.md)
- [Webhooks & Events](/evaluation/docs/integrations/webhooks.md)
- [Custom Pipelines](/evaluation/docs/integrations/custom-pipelines.md)
- [Security & Compliance](/evaluation/docs/security-compliance/overview.md)
- [PHI Handling](/evaluation/docs/security-compliance/phi-handling.md)
- [Encryption](/evaluation/docs/security-compliance/encryption.md)
- [Access Logging](/evaluation/docs/security-compliance/access-logging.md)
- [Data Retention & Deletion](/evaluation/docs/security-compliance/data-retention.md)
- [Voice AI Overview](/evaluation/docs/voice/overview.md)
- [Patient Triage](/evaluation/docs/voice/triage.md)
- [Transcripts & Audio](/evaluation/docs/voice/transcripts.md)
- [DICOM & Medical Imaging](/evaluation/docs/images/dicom.md)
- [Pathology (WSI)](/evaluation/docs/images/pathology.md)
- [Clinical Notes Overview](/evaluation/docs/notes/overview.md)
- [Clinical Documentation](/evaluation/docs/notes/clinical.md)
- [Clinician Review](/evaluation/docs/workflows/clinician-review.md)
- [AI / ML Metrics Definitions](/evaluation/docs/glossary-appendix/ai-ml-metrics.md)- [Regulatory References](/evaluation/docs/glossary-appendix/regulatory-references.md)
- [Evaluation Rubrics Library](/evaluation/docs/glossary-appendix/rubrics-library.md)
- [Versioning & Stability](/evaluation/api-reference/overview/versioning.md)
- [Request Format](/evaluation/api-reference/overview/request-format.md)
- [Pagination & Filtering](/evaluation/api-reference/overview/pagination.md)
- [API Keys](/evaluation/api-reference/authentication/api-keys.md)
- [OAuth 2.0 / SSO](/evaluation/api-reference/authentication/oauth.md)
- [Scopes & Permissions](/evaluation/api-reference/authentication/scopes.md)
- [Role-Based Access Control](/evaluation/api-reference/authentication/rbac.md)
- [Org & Project Permissions](/evaluation/api-reference/authentication/org-project-permissions.md)
- [Audit Logs](/evaluation/api-reference/authentication/audit-logs.md)
- [Error Format](/evaluation/api-reference/errors/overview.md)
- [Error Code Reference](/evaluation/api-reference/errors/codes.md)
- [Retry & Backoff Strategy](/evaluation/api-reference/errors/retry.md)
- [Rate Limits](/evaluation/api-reference/rate-limits/overview.md)
- [Headers & Monitoring](/evaluation/api-reference/rate-limits/headers.md)
- [Create Dataset](/evaluation/api-reference/datasets/create.md)
- [List Datasets](/evaluation/api-reference/datasets/list.md)
- [Get Dataset](/evaluation/api-reference/datasets/get.md)
- [Update Dataset](/evaluation/api-reference/datasets/update.md)
- [Delete Dataset](/evaluation/api-reference/datasets/delete.md)
- [Dataset APIs](/evaluation/api-reference/datasets/dataset-apis.md)
- [Evaluation APIs](/evaluation/api-reference/datasets/evaluation-apis.md)
- [Scoring & Metrics APIs](/evaluation/api-reference/datasets/scoring-metrics-apis.md)
- [Human Review APIs](/evaluation/api-reference/datasets/human-review-apis.md)
- [Audit & Provenance APIs](/evaluation/api-reference/datasets/audit-provenance-apis.md)
- [Errors & Status Codes](/evaluation/api-reference/datasets/errors-status-codes.md)
- [Rate Limits & Quotas](/evaluation/api-reference/datasets/rate-limits.md)
- [Samples Overview](/evaluation/api-reference/samples/overview.md)
- [Create Sample](/evaluation/api-reference/samples/create.md)
- [Batch Upload](/evaluation/api-reference/samples/batch.md)
- [List Samples](/evaluation/api-reference/samples/list.md)
- [Get Sample](/evaluation/api-reference/samples/get.md)
- [Schemas by Modality](/evaluation/api-reference/samples/schemas.md)
- [Models Overview](/evaluation/api-reference/models/overview.md)
- [Register Model](/evaluation/api-reference/models/register.md)
- [List Models](/evaluation/api-reference/models/list.md)
- [Get Model](/evaluation/api-reference/models/get.md)
- [Model Versions](/evaluation/api-reference/models/versions.md)
- [Create Evaluation](/evaluation/api-reference/evaluations/create.md)
- [List Evaluations](/evaluation/api-reference/evaluations/list.md)
- [Get Evaluation](/evaluation/api-reference/evaluations/get.md)
- [Get Evaluation Status](/evaluation/api-reference/evaluations/status.md)
- [Cancel Evaluation](/evaluation/api-reference/evaluations/cancel.md)
- [Compare Evaluations](/evaluation/api-reference/evaluations/compare.md)
- [Tasks Overview](/evaluation/api-reference/tasks/overview.md)
- [Scores Overview](/evaluation/api-reference/scores/overview.md)
- [Reviewers Overview](/evaluation/api-reference/reviewers/overview.md)
- [Reviews Overview](/evaluation/api-reference/reviews/overview.md)
- [Webhooks Overview](/evaluation/api-reference/webhooks/overview.md)
- [SDKs Overview](/evaluation/api-reference/sdks/overview.md)

## Expert Review

- [Bring your experts](/talent/introduction.md)
- [RLHF evaluations](/talent/guidance/rlhf-evaluations.md)
- [Domain writing for RL environments](/talent/guidance/domain-writing.md)
- [Create your account](/talent/how-to/create-account.md)
- [Setup your profile](/talent/how-to/setup-profile.md)
- [Accept project work](/talent/how-to/apply.md)
- [Setup payments](/talent/how-to/payments.md)
- [Invite additional specialists](/talent/how-to/referrals.md)
- [Getting started](/talent/working/getting-started.md)
- [Project onboarding](/talent/working/onboarding.md)
- [Project offboarding](/talent/working/offboarding.md)
- [Policies & compliance](/talent/policies/overview.md)
- [FAQ & support](/talent/support/faq.md)

## OpenAPI specs

- [Control plane PDP OpenAPI](https://raw.githubusercontent.com/Akhara-AI/akhara-docs/main/control-plane/api-reference/openapi.json): Policy Decision Point API (`api.akhara.dev`): authorize, agents, evidence
- Repo path: `control-plane/api-reference/openapi.json` (also referenced from `docs.json` → `api.openapi`)

## Public API hosts

| Product | Host | Notes |
|---|---|---|
| Environments eval API | `https://agi.akhara.ai` | `POST /evaluations/run`, dashboard APIs |
| Environments admin | `https://admin-agi.akhara.ai` | Run / trajectory UI |
| Control plane PDP | `https://api.akhara.dev` | Decisions only; never executes tools |
| Control plane console | `https://console.akhara.dev` | Workspace / keys |
| Evaluation app / API | `https://app.akhara.ai`, `https://api.akhara.ai` | Hosted evals + scoring APIs |
| Expert Review portal | `https://talent.akhara.ai` | Enterprise-invited expert workflows (docs path `talent/`) |

## Optional

- [Full docs export](/llms-full.txt): Concatenated Markdown of all pages (Mintlify-generated when hosted)
- Source repository: https://github.com/Akhara-AI/akhara-docs