e-volv

The register

39 capabilities in ten groups. Anything that does not reach every workspace today carries its code in the margin, so you can read this page once and know what you would actually get.

Specialist agents that ship code, not just suggestions

Five base agent types with curated toolkits, memory, multi-agent coordination, a signed identity per agent, and replayable evals, all running in an isolated Python service.

  1. Base agent typesSenior Developer, Code Reviewer, Tech Writer, Product Manager, and fully Custom agents.47 built-in tools
  2. Agent memoryPersistent memory tools let agents recall conventions, recurring issues, and decisions.
  3. Multi-agent sessionsA coordinator delegates work across specialist agents and synthesizes the result.
  4. MCP serversConnect remote MCP servers over streamable HTTP or SSE and their tools appear namespaced in every agent run.
  5. Agent identity & ledgerOpt-inEach agent gets a stable identity, a signed short-lived token, and a ledger of everything it touched.Opt in: requires the agents.identity flag
  6. Eval harnessOpt-inReplay any past step against a new prompt or model with write tools stripped, then diff what changed.Opt in: requires the agents.eval flag

Visual workflows that run your delivery pipeline

Drag, drop, and connect triggers and actions on a square-grid canvas. Publish, version, and execute with full observability.

  1. Visual builderReact Flow canvas with trigger, agent, git, CI, ticket, and approval nodes aligned to the 22px grid.17 node types
  2. Ticket nodesCreate, update, comment, read, and list tickets as first-class workflow steps across Jira, Linear, and ClickUp.5 ticket node types
  3. Reusable blocksSave a group of steps as a workflow block and reuse it across workflows.
  4. Dry-run testingRun the whole graph against sample trigger data with side effects stubbed out.
  5. AI generationGenerate a complete workflow (nodes, edges, and config) from a plain-English prompt.
  6. Execution engineTopological scheduling, parallel branches, step retries, and live SSE streams.
  7. Template engineReference trigger fields and previous step outputs with {{trigger.*}} and {{outputs.*}}.

React to every event in your toolchain

Webhooks from 7 providers, scheduled cron, manual runs, slash commands, and @mentions, all normalized to a canonical event.

  1. Webhook normalizationGitHub, GitLab, Bitbucket, Azure DevOps, Jira, Linear, and ClickUp payloads become NormalizedEvent.7 providers
  2. Trigger conditions11 condition operators let you filter events before execution starts.
  3. Scheduled & manualSchedule recurring checks or trigger workflows manually from the dashboard.

Connect the tools your team already uses

Git providers, ticket systems, LLM providers, MCP servers, a signed MCP registry, encrypted vault storage, and a scoped public API in one hub.

  1. Git providersGitHub, GitLab, Bitbucket, and Azure DevOps adapters with PR, issue, branch, and CI support.
  2. Ticket systemsJira, Linear, and ClickUp adapters for reading, creating, and updating tickets.
  3. Ticket boardsConnect several boards per tracker and read them as a single kanban, filtered, searchable, and writable.
  4. LLM modelsOpenAI, Anthropic, Azure OpenAI, AWS Bedrock, OpenRouter, and DeepInfra.
  5. VaultAES-256-GCM encrypted workspace secrets, never returned in list endpoints.
  6. MCP registryOpt-inVersioned, signed MCP server manifests that agents reference by name and semver range.Opt in: requires the mcp.registry flag
  7. Public APIOpt-inScoped API keys open a standalone surface: start workflows from your own systems, ingest runs, post heartbeats, read metrics, rotate secrets, publish manifests.Opt in: most scopes need a per-workspace flag17 API scopes

Security, audit, and control at scale

JWT auth, workspace RBAC, encrypted vault with scheduled secret rotation, audit logs, token budgets, and isolated sandbox execution.

  1. SecurityJWT auth, bcrypt hashing, webhook signature verification, and service-token guards.
  2. AuditAppend-only audit events for auth, vault, integrations, triggers, and workspaces.
  3. BudgetMonthly token budgets with hard-stop and warning thresholds, enforced pre-execution.
  4. Approval gatesPause any workflow at an approval node until someone signs off, or the request expires.
  5. SandboxRun tools in Docker containers or Cloud Run sandboxes with limited CPU, memory, and network.
  6. Secret rotationOpt-inRotation policies with due dates, verification against the real provider, and a grace window where old and new both work.Opt in: requires the vault.rotation flag

A map of the code, not a folder tree

Every symbol, every call and import, ranked by how central it is, so agents retrieve by meaning and reviewers see what a change actually touches.

  1. Code MapA tree-sitter index of every symbol and the edges between them, ranked by PageRank and browsable from components down to individual functions.3 languages parsed
  2. Semantic code searchAgents find code by describing what it does. Retrieval is by embedding similarity, ranked by the structure of the graph.1024 embedding dimensions

Observe every workflow, token, and failure

KPI dashboards, execution volume, token trends, agent leaderboards, tool usage, failure analysis, and cost estimation.

  1. Analytics dashboardRecharts-based dashboards for execution volume, token consumption, and agent performance.7 analytics endpoints
  2. Execution viewerSSE streaming, step-level retry, and full input/output inspection.

Know when automation stops working

Late-run detection for scheduled jobs, alert rules on failure metrics, and SLOs with error budgets, for e-volv workflows and for anything else you point at it.

  1. Scheduled-job monitoringOpt-inEvery scheduled trigger is watched for lateness and hung runs; external jobs check in with one curl.Opt in: requires the monitoring.scheduled flag
  2. Alert rules & metricsOpt-inThreshold rules on success rate, tool failures, failed runs, and error-budget burn; a Prometheus endpoint for the rest.Opt in: requires the alerts.rules flag8 alert metrics
  3. SLOs & error budgetsOpt-in"95% of Code Review runs succeed over 30 days", the budget left, multi-window burn rate, and a README badge.Opt in: requires the slo.tracking flag5 SLO indicators

Logs, traces, and the failure graph that explains them

Two SDKs, OTLP ingest, TimescaleDB storage with a small query language, tail, dashboards and error groups, and an incident builder that names what failed first, how far it spread, and why.

  1. Log capture and searchOTLP ingest from any OpenTelemetry client or the JS and Python SDKs, stored in TimescaleDB with a query language, live tail, dashboards and error groups.30 days hot retention
  2. Incidents and the failure graphErrors are extracted as they land, fingerprinted into groups, and when they burst the incident builder walks failing traces and the call graph to produce a failure graph an agent then explains.180 days of traces & incidents

Production-ready workflow templates

Import pre-built playbooks in one click. Each playbook creates the agents, workflows, and triggers you need for a common SDLC task.

  1. Playbook catalog89 documented playbooks across 10 categories: review, docs, ci, security, planning, interaction, observability, performance, agent ops, compliance.89 playbooks

Read one against your own repository

Every capability above is reachable from a free workspace except the ones carrying a code, and those say so where they are listed rather than in a footnote.

Start freeBrowse the catalogue

We use only essential cookies to keep you signed in and the app working. With your consent we may enable optional analytics to help us improve e-volv. See our Privacy Policy and Terms.