Skip to content

Latest commit

 

History

History
96 lines (82 loc) · 7.38 KB

File metadata and controls

96 lines (82 loc) · 7.38 KB

Unique Value Proposition

Statement

The only platform that lets any developer visually design, locally execute, and git-commit multi-model multi-agent workflows that run identically in their IDE, terminal, desktop canvas, and CI/CD pipeline — no cloud dependency, no vendor lock-in, no framework expertise required.

Proof Points

  1. Every competitor owns at most two surfaces; Relavium owns all four (desktop, VS Code, CLI, web). The identical engine runs on the three Phase-1 execution surfaces — desktop, VS Code, and CLI — so a workflow behaves the same on each; the Phase-2 web portal is a browser control-plane surface (usage, quota, governance), not a fourth identical-engine runtime.
  2. No chat-driven assistant lets you keep — and ship — the conversation. Relavium agent sessions are persistent, resumable, and exportable: auto-saved to encrypted local history, resumable on any surface, and one-click exportable to a reviewable .relavium.yaml scaffold. Competitors' sessions evaporate; Relavium's become committable, re-runnable workflows. See decisions/0024-agent-first-entry-point-agentsession.md and decisions/0026-session-export-to-workflow.md.
  3. No competitor produces version-controllable workflow artifacts. Relavium workflows are first-class git objects (.relavium.yaml in .relavium/).
  4. **No competitor combines visual design + local execution + multi-model routing
    • multi-agent orchestration** in a single product.
  5. Workflow portability means a workflow built by one developer runs unchanged in CI/CD — closing the gap between local AI use and automated pipelines that every other tool leaves open.
  6. BYOK-local mode means zero data leaves the machine — guaranteed, permanently. In Relavium's first-class BYOK-local ("Private mode"), runs execute on the user's machine and LLM calls go straight to the providers under the user's own keys; nothing transits Relavium. This is a mode-scoped guarantee kept permanently non-degraded — a trust and compliance advantage no SaaS-first competitor can match. (Phase 2 adds an opt-in managed-inference convenience mode that proxies LLM egress through Relavium's keys; BYOK-local always stays available for users who want the zero-egress guarantee. See product-constraints.md and decisions/0012-managed-inference-dual-mode.md.)
  7. Convenience without lock-in (Phase 2, opt-in). Managed inference lets users start with zero key setup, while BYOK-local remains a one-switch escape hatch — same product, your key — so the convenience never becomes lock-in.

Positioning

Relavium is two entry points on one engine: a conversational coding agent and a workflow-orchestration layer above the editor — not another editor. You start in chat and graduate to a committed workflow; the same engine backs both. The five properties that, taken together, no one else offers:

Property What it means Who lacks it
Chat-first Start in a conversational coding session on any surface (relavium chat, a desktop Chat tab, a VS Code panel) that is persistent and one-click exportable to a workflow Claude Code/Cursor/Cline (chat is the only surface, no export path), CrewAI/AutoGen (no conversational entry)
Visual Drag-and-drop canvas with live execution on the node face CrewAI/AutoGen (code only), Claude Code/Cursor/Cline (chat only)
Git-native Workflows are diffable, PR-able YAML files n8n (proprietary JSON), Claude Code/Cursor (no persistence)
Multi-model Per-agent model choice + fallback chains across providers Claude Code (Claude only), Copilot Agent (MS/OpenAI only)
No lock-in Runs in standard VS Code, any terminal, any git host, locally Cursor (fork lock-in), Copilot Agent (GitHub lock-in)

Competitive Framing

A condensed read of the per-competitor strengths and the gaps Relavium fills — this table is the surviving canonical competitive framing. (The dated competitive-landscape analyses it distilled were removed on 2026-06-10.)

Competitor Their strength The gap Relavium fills
Claude Code Deep Claude integration, strong CLI + editor UX, conversational coding agent Single-model, single-agent, sessions are ephemeral with no export path, no visual canvas, no version-controllable workflows, no CI/CD path. Relavium meets it head-on with its own first-class chat agent, then adds the missing exit: export the session to a committed, multi-model, re-runnable workflow when users outgrow single-agent/single-model.
Cursor Best-in-class AI-native editor Editor-only, fork lock-in, no workflow abstraction, no reuse/sharing. Relavium sits above any editor, keeping the user's preferred tools.
Continue.dev Open-source, multi-model, BYO keys Stops at chat/autocomplete — no orchestration, no workflow files, no CLI, no desktop. Natural upgrade for users who want to chain and persist agents.
Cline / Roo-code Aggressive agentic execution (fs, shell, browser) Every run is ephemeral and manual; no reuse, no team sharing, no canvas. Relavium makes agentic runs reusable and composable.
GitHub Copilot Agent Mode Massive GitHub distribution, enterprise trust GitHub/MS ecosystem lock-in, no multi-model, no local-first, no portable workflow files. Relavium is VCS-, model-, and surface-agnostic.
n8n / Zapier Mature visual builders, huge connector libraries Not AI-native, no developer surfaces (no VS Code/CLI), no local-first, no code/repo context. Relavium is the developer-native, AI-first equivalent.
CrewAI / AutoGen Powerful programmable multi-agent topologies Python-only, no IDE/canvas, requires code to define agents, no visual artifact. Relavium delivers the same power via a canvas + readable YAML, no framework expertise.

Two Entry Points, One Wedge

Relavium sits at both ends of a developer's flow. Chat is the hook: a familiar, zero-setup conversational coding agent — the same surface developers already reach for — is how they arrive, and relavium chat is the fastest path to first value. The workflow file is the wedge: because a workflow is a portable, readable YAML file, the workflow file itself becomes the viral mechanic — a developer shares a .relavium.yaml, a teammate runs it unchanged in their own editor, terminal, or CI. The chat-to-workflow continuum is what connects them: the moment a session proves a flow, exporting it to a committed workflow turns a private conversation into shareable team infrastructure. Value scales with team size — the workflow file is team infrastructure, not a personal preference. This positioning underpins the go-to-market and pricing strategy (gate on collaboration and scale, not on capability).

How this maps to adoption per segment — execution mode, key model, tier, and the upgrade path for individuals, small teams, and enterprises — is detailed in deployment-models.md.