Deterministic Assistant
A scripted flow you control node-by-node — the same logic, now editable on the canvas with versioning and live simulation.
✣migrate off your legacy bot
Your legacy bot is stuck in 2019.
Make it an AI agent in a day.
Upload your existing bot's export. Convoship auto-detects it and converts your intents, entities, and flows into a deterministic Assistant or an autonomous AI Agent — no rebuild. No bot yet? Sketch one and we'll build it.
Enterprise-ready on day one — row-level isolation, MFA, an immutable audit log, and encrypted secrets in every workspace, not a future tier.
Bring any bot export·Deterministic or agentic·Cross-session memory·Per-node heatmap
migrate, don't rebuild
Bring the bot you already run.
Most platforms only let you build new. Convoship migrates you off the one you've outgrown. Upload your bot's export and we auto-detect its structure from the archive — no manual mapping — then convert intents, entities, and flows into Convoship's canonical schema. On a platform we don't parse yet? We add it.
Intents
every intent and its training phrases, matched and routed
Entities
slots and entity types, mapped into the session
Flows
decisions, branches, and fulfillment, preserved
Tools
API calls and webhooks, kept as real tool nodes
your export, materialized as
Autonomous AI Agent
An agentic runtime with mission, tools, persona, and guardrails — for reasoning over the long tail, not just scripted paths.
Single model pass with an offline heuristic fallback — imports work even without a provider key.
no bot to bring? draw it
Or sketch the one you want.
Starting from scratch? Draw the conversation the way your team already whiteboards it — or drop in a doc you already have. A deterministic parser plus a single model pass turn it into a runnable Assistant or AI Agent on the same canvas your migrations land on.
Draw.io
boxes, arrows, decision diamonds, service calls
PDF / Word
a runbook, spec, or support-macro doc
Image / photo
a snapshot of the whiteboard, read by vision
Blank canvas
start from one node and build it live
Same canonical schema, same editor, same deterministic-or-agentic output as a migration — just a different on-ramp.
how convoship works
From your old bot to a production AI agent — in three moves.
A deterministic parser reads your existing bot's export — or a Draw.io sketch — into a typed baseline. A single model pass cleans labels, infers slots, and lifts service calls. You review the result on a familiar visual canvas.
apply_loantenure?amount?credit_check
Import your bot
Upload your existing bot's export — Convoship auto-detects its structure from the archive itself, no manual mapping. (No bot? Draw the flow in Draw.io.) Intents, entities, decisions, and service calls all become structured flow primitives.
Claude Sonnet 4.6
anthropic · claude api
✓12 nodes
✓14 edges
✓4 labels cleaned
✓2 slots inferred
412ms · 1.2k tokens · $0.004
Convoship converts it
Auto-detect the source, clean labels, infer entity slots, and wrap HTTP nodes as real tool calls with auth, retries, and response mapping — then materialize it as a deterministic Assistant or an AI Agent, your choice.
<script src=
"https://embed.convoship.org/v1.js">
</script>
ConvoshipQueue.push(['init', agent])
Edit, improve, embed
Tune prompts and decisions on the canvas, simulate live, watch the per-node drop-off heatmap, then paste a one-line script tag on any website.
try it yourself
Sketch it. Watch it run.
No signup, no sandbox to configure. Walk the exact path your team will: drop a diagram, watch Convoship parse it, then talk to the agent it produced — runtime trace and all.
loan-application.drawio
apply_loanamount?tenure?credit_checkyour whiteboard
Deterministic parser first, then one Claude pass to clean it up.
built for the CX scorecard
The numbers your board asks about — measured for you.
Every conversation rolls up into the metrics a support leader is judged on. Not vanity charts: resolution, satisfaction, handle time, and dollars saved — each trended against the prior period, on every agent.
Resolution rate
Resolved without a human
True containment — conversations that never escalated to a person.
CSAT
👍 / 👎 after the chat
Customer-reported satisfaction across rated conversations.
Avg handle time
Minutes per conversation
How long a customer spends to reach a resolution.
Cost-per-contact saved
Deflected × your cost
Set your blended cost-per-contact; watch deflection turn into dollars.
product tour
Migrate it, ship it, and watch it improve.
Import your legacy bot (or a sketch), ship it as a deterministic Assistant or an autonomous AI Agent, give it cross-session memory, then watch a per-node heatmap show you where to improve — all in one workspace. No tab-juggling, no second tool.
Migrate any bot
Import your existing bot's export — auto-detected and converted, no manual rebuild
Assistant or AI Agent
Ship scripted-deterministic or agentic — mission, tools, persona, guardrails
Cross-session memory
Agents recall a returning user's context and top use-cases
Drop-off heatmap
Per-node abandonment on the canvas, so you fix the weak spots
Visual builder
React Flow canvas with library and inspector
Live simulator
Send messages, inspect context, and replay history
Channels
Embed widget, voice, WhatsApp, and webhook
Workflows
Chain triggers, assistants, agents, decisions, and webhooks
app.convoship.org/agents/loan-app/builder
Intent · Entry
apply_loan
12 phrases · 2 params
Entity
Capture amount
@number -> loan_amount
Tool
credit_check
POST /credit/check
Script
Compose reply
templates {{rate}}
built for builders
The whole platform, in one screen.
Everything reads the same canonical conversation-flow schema — legacy-bot migrations, sketch imports, scripted Assistants, and agentic AI Agents alike. Migrate, import, edit, simulate, deploy, and measure all share one JSON.
Migrate from legacy platforms
Your existing bot's export — auto-detected from the archive and converted to a deterministic Assistant or an AI Agent. On a format we don't parse yet? We add it. A heuristic fallback means imports work even offline.
Multi-source sketch import
Draw.io, PDF, Word, and image uploads. One shared flow schema with source-specific parsers.
Visual flow editor
React Flow canvas with library and inspector. Intents are first-class. Subflows compose.
Sandboxed code nodes
Python with session-scoped variables. No more glue scripts living elsewhere.
Real tool integration
HTTP nodes with auth, templated body, JSONPath output mapping, retries, and encrypted secrets.
Entity-aware slot filling
System types for numbers, dates, emails, and durations. Custom list entities with synonyms.
Embed anywhere
Small IIFE bundle, one script tag, queue snippet, tunable theme, lifecycle hooks, public API.
Built-in analytics
Conversation counts, session health, latency, fallback rates, completion, and audit history.
Live simulator + debug
Open any conversation history in the simulator and inspect the full runtime context in a popup.
Production foundations
Workspace auth, MFA, audit log, secrets vault, versioned drafts, and usage metering.
Deterministic Assistant or AI Agent
Same import, your choice of output: a scripted, deterministic Assistant or an autonomous AI Agent with LLM reasoning, tool calls, guardrails, evals, and spend caps. Upgrade an Assistant to an Agent in one click.
Cross-session memory
AI Agents recall a returning user's prior context and top use-cases, so conversations pick up where they left off instead of starting cold.
Drop-off heatmap + improvement loop
Live per-node drop-off overlaid on the canvas. Pair it with the AI flow doctor, persona-simulation stress tests, and visual version diff to close the loop.
Workflow chains
Orchestrate assistants, agentic agents, webhooks, and branches — e.g. qualify with an agent, hand off to a scripted intake flow.
ships better every week
The only builder that shows you where to improve.
Shipping is the start, not the finish. Convoship turns live traffic into your next fix — a per-node drop-off heatmap right on the canvas, plus the tools to act on it.
Drop-off heatmap
Every node shows the share of sessions that abandon there, overlaid on the flow you edit.
AI flow doctor
A pre-publish gate that catches dead ends, unreachable nodes, and missing slots before they ship.
Persona simulation
Stress-test against synthetic customer personas and find where the agent struggles before customers do.
Visual version diff
Compare any two versions node-by-node, so every change is deliberate and reviewable.
Auto-detected from the export
any bot
no manual mapping or rebuild
Import to working agent
minutes
auto-detected · no manual mapping
Conversion fidelity (node F1)
>= 0.99
CI gate · 30 gold flows
Cost per import
$0.004
one bounded model pass
under the hood
One schema. Every surface reads it.
The importer, runtime engine, embed SDK, and eval harness all share a single canonical conversation-flow schema. Export it, version it, and migrate it across workspaces.
loan-application.drawioXML
<mxGraphModel>
<root>
<mxCell id="n1" type="intent" value="apply_loan"/>
<mxCell id="n2" type="decision" value="amount > 50000?"/>
<mxCell id="n3" type="api" value="credit_check"/>
</root>
</mxGraphModel>agents/loan-app/v8.jsonJSON
{
"id": "agt_loan-app",
"version": 8,
"intents": [{
"name": "apply_loan",
"phrases": 12,
"parameters": ["loan_amount"]
}]
}~/acme-bank
Studio · Deploy · loan-application
-> import · drawio · claude-sonnet-4-6 (anthropic) · 412ms · $0.004
-> validate · 12 nodes · 14 edges · 0 errors
-> publish · version 8 · simulate ok
-> token · cdp_8c2...ke7 · embed snippet ready
✓ live on acme.bank · 42 sessions today
how it stays correct
Deterministic where it matters, LLM where it helps.
Convoship doesn't ask the LLM to invent your conversation. A deterministic parser turns a diagram — or a legacy bot's export — into structured nodes; a single Claude pass cleans labels, infers slots, and lifts service calls. A validator + auto-repair pass catches the trivial stuff before any LLM round-trip — so the only thing the model is asked to do is the part where models are actually good.
- 1
Deterministic parser — Draw.io / PDF / Word / image / legacy export
Pure code, no LLM. Auto-detects the source (a sketch, or an existing bot's export) and reads it into a typed AST of nodes, edges, and labels.
- 2
Validator + auto-repair
Rule-based pass fixes missing entries, dangling edges, and duplicate ids before validation. Trivial issues never consume a model call.
- 3
Single Claude pass (claude-sonnet-4-6)
Cleans labels, infers entity slots, lifts service calls. Cost is bounded per import; output is constrained to the canonical schema.
- 4
Validator again, then commit
Any leftover errors are surfaced as actionable issues on the canvas, not silently shipped. Drafts are versioned; publish is explicit.
- 5
Gold-eval regression suite
Thirty real-world gold flows gate the extractor in CI. Node F1 ≥ 0.99 is a release blocker, not a best-effort metric.
why convoship
What other platforms can't do.
Plenty of tools build a new bot. Convoship is the one that takes the bot you already run, makes it deterministic or agentic, gives it memory, and shows you where to improve.
Migrate off, not just build on
Bring any existing bot's export off the platforms that only let you build new and never leave — and we add new formats on request.
Deterministic and agentic, one workspace
Ship a scripted Assistant or an autonomous AI Agent from the same import — and upgrade one to the other in a click.
Memory and self-improvement built in
Cross-session memory plus a per-node heatmap and flow doctor — not a separate analytics add-on.
Honest, measured metrics
Every number — node F1 ≥ 0.99, ~500ms extraction — comes from CI eval gates, not a pitch deck.
Enterprise controls on day one
Row-level isolation, MFA, immutable audit logs, and encrypted secrets in every workspace — no add-on tier.
why we built convoship
We were tired of legacy bots no one could leave.
Every team we talked to was paying for a conversational bot they'd outgrown — a rigid, forgetful build from years ago that was seemingly impossible to leave. (The rest had the flow mapped on a whiteboard and never shipped it.) Convoship exists to close that gap: the bot you already run becomes a reasoning, remembering AI Agent in an afternoon, not a quarter — and you finally get off the platform you've outgrown.
Fayaz · Founder, Convoship
Deterministic first, LLM second
We don't ask a model to invent your conversation. A real parser does the structural work; the model only does the part models are actually good at. That's why imports are cheap, fast, and repeatable.
Honest metrics, in public
Every number on this page — latency, cost per import, node F1 — comes from our eval harness and CI gates, not a pitch deck. If a metric isn't measured, we don't claim it.
Enterprise controls on day one
Workspace isolation, MFA, audit logging, and encrypted secrets aren't a future enterprise tier. They're in every workspace, because we'd want them before trusting a vendor too.
chat widget
One line of script. Any website.
The embed SDK renders the same runtime your team tested in Studio, so what passes in the simulator is the behavior your customers see.
acme.bank
ACME BANK
Personal loans, on your terms.
Rates from 6.9% APR. Apply in five minutes. No hidden fees.
Apply now
Loan Application
How much would you like to borrow?
25 thousand
solutions
Built for the conversations your team already runs.
Convoship works wherever a team has already mapped the conversation — on a whiteboard, in a runbook, or in a bot you already run. Sketch it or migrate it; the same path serves regulated industries and consumer-grade self-service.
Banking
Onboarding, lending, and account servicing
Capture KYC, qualify loan applications, and answer balance / transaction questions. Decision diamonds gate credit checks; service calls reach core banking APIs through HTTP nodes with encrypted secrets, retries, and JSONPath output mapping.
- KYC slot-filling
- Credit decisioning
- Audit log + MFA
Retail
Order status, returns, and product search
Pull live order data from your OMS, surface return windows, and route shoppers to PDPs that actually match what they described. Knowledge collections cover policies; intents cover transactional tasks.
- OMS integration
- RAG for policy lookup
- Web + WhatsApp deploy
Healthcare
Triage, scheduling, and benefits questions
Run pre-visit triage flows, book or reschedule appointments through EHR APIs, and explain benefits without exposing PHI to the LLM. Guardrails restrict outputs to approved language; secrets stay vaulted.
- PHI-safe prompts
- EHR API tool calls
- Custom guardrails
Hospitality
Reservations, concierge, and upsell
Take bookings, recommend room upgrades, and answer property questions across web, voice, and WhatsApp from the same agent. Localized copy per intent; analytics expose conversion by funnel step.
- Voice + chat parity
- Localized intents
- Funnel analytics
Customer support
Deflection and live-agent handoff
Resolve top-N FAQs with knowledge collections, run intent-driven self-service for the next tier, then hand off to a human with full session context. Workflow chains coordinate AI Agent reasoning with scripted intake.
- Knowledge (RAG)
- Workflow chain handoff
- Live transcript export
IT helpdesk
Ticket triage and self-serve fixes
Categorize incoming tickets, walk users through known runbook fixes, and open a ticket in your ITSM with the full diagnosis when the bot can't resolve it. Tools call Jira / ServiceNow / Zendesk under workspace secrets.
- Runbook flows
- ITSM tool calls
- Workspace audit log
security & compliance
Production foundations on day one.
Convoship was built for teams whose security review never gets skipped. Every workspace ships with the controls your auditors expect — no add-on tier, no add-on price.
Identity & access
Workspace roles (owner, admin, developer, editor, viewer), MFA enforcement, refresh-token rotation, configurable session lockout, and a workspace-wide revoke-all-sessions action.
Encrypted secrets
Workspace secrets vault with envelope encryption (Fernet today, KMS-ready). Tool nodes reference secrets by name — credentials never enter prompts, exports, or logs.
Row-level security
Postgres RLS enforces workspace isolation on every query. The app role cannot bypass RLS; cross-workspace data exposure is structurally impossible, not just policy-enforced.
Audit trail
Every mutation — agent edits, deployments, secret reads, member role changes — lands in an immutable audit log. Filter by actor, action, target, and time range — the audit evidence your security review will ask for.
Usage metering & budgets
Daily LLM spend caps per AI Agent, per-workspace conversation counts, and Prometheus metrics for runtime sessions, turns, and tool calls. No surprise bills, no silent failures.
Sandboxed code execution
Python nodes run with a strict per-node timeout and session-scoped variables only. No filesystem, no outbound network unless explicitly proxied through a tool node.
founding design partners
We're taking on a small group of founding customers.
Convoship is new, and we're deliberate about our first enterprise deployments. Instead of a wall of logos we haven't earned yet, we're inviting a limited cohort of design partners to build with us directly — and lock in founding terms.
Limited cohort. We'd rather serve a few teams exceptionally than many poorly.
- A direct line to the founder and engineering — not a ticket queue
- Hands-on onboarding: we help import and ship your first agents with you
- Real influence on the roadmap — your use cases shape what we build next
- Founding pricing, locked for the life of the partnership
- A named security and architecture review before you commit
Built for enterprise teams.
Convoship is sold on annual agreements with the security, scale, and support your organization expects. Every workspace ships on the Enterprise plan — no agent count caps on imports or sessions.
Enterprise
Customannual agreement
For regulated industries and high-volume conversational programs.
- Unlimited agents, imports, and runtime sessions
- Legacy bot migration — bring any export
- Deterministic Assistants and agentic AI Agents
- Cross-session memory + per-node drop-off heatmap
- Workspace roles, MFA, and audit log
- Encrypted workspace secrets vault
- WhatsApp, voice, webhooks, and web embed
- Knowledge (RAG), workflows, and dedicated support
enterprise faq
The questions your security and procurement teams will ask.
Straight answers, including the ones most vendors dodge. If something here matters to your evaluation, raise it on the call — we'd rather over-explain than oversell.
Can we migrate our existing bot?
Yes — that's a first-class path, not an afterthought. Upload your export ZIP and Convoship auto-detects its structure by inspecting the archive itself, parses the intents, entities, and flow, and lets you materialize it as either a deterministic scripted Assistant or an autonomous AI Agent. We support the major conversational platforms today and add new export formats on request. Enrichment uses a single model pass with an offline heuristic fallback, so imports work even without a model-provider key.
Where does our data go, and which AI provider do you use?
Conversation extraction and runtime reasoning use Anthropic's Claude API (Sonnet for extraction, Haiku for latency-sensitive paths). Your agent configuration and conversation data live in your workspace, isolated at the database level by Postgres row-level security. We don't sell data, and we don't use your data to train models.
Are you SOC 2, ISO 27001, or HIPAA certified?
Not yet — we're a new platform and we won't claim a certification we don't hold. What we can show you today is the actual control set: row-level tenant isolation, MFA, immutable audit logs, envelope-encrypted secrets, and sandboxed code execution. We're glad to walk your security team through our architecture and our roadmap toward formal attestation. Full detail is on our Trust page.
Can we self-host or deploy in our own VPC?
Yes. Convoship runs as a containerized stack (FastAPI, Postgres, Redis) and supports self-hosted and VPC deployments, including self-hosted vision endpoints for on-prem import. Cloud, self-hosted, and hybrid are all on the table for design partners.
How do you handle secrets and third-party credentials?
Tool integrations reference secrets by name from a workspace vault with envelope encryption (KMS-ready). Credentials never enter prompts, exports, or logs, and the application database role cannot bypass workspace isolation.
What about PHI and other regulated data?
Guardrails can restrict agent outputs to approved language, and you can design flows that keep sensitive values out of LLM prompts — capture them and pass them through tool calls instead. We review your specific compliance requirements during onboarding rather than wave them away.
What does pricing look like?
Convoship is sold as an annual enterprise agreement with no caps on agents, imports, or runtime sessions. Founding design partners lock in founding pricing. Talk to us about your volume and deployment model.
What support and SLA do we get?
Design partners get a direct line to the founder and engineering, hands-on onboarding, and a response commitment we'll put in writing in your agreement. Formal SLA tiers will follow as we scale.
✣ migrate any bot -> AI agent
Get off your legacy bot.
Ship an AI agent this week.
See your own bot's export become a production AI agent in a single working session — reasoning, cross-session memory, and a drop-off heatmap included.