NEWAgentic AI v3 — 14 tools that actually do the work. Read the launch →
AI-native project management

Project management,
finally native to agentic AI.

Seshira's built-in agent doesn't just answer questions — it executes. 14 tools that create sprints, move cards, assign owners, and detect blockers. Jira features, Linear polish, AI that actually ships work.

Free for 10 users 100 AI credits/mo No credit card
morphlix/Seshira/Mobile App · Sprint 22
Live · 5 online
Sprint 22 · Mobile App14d · 41/44 pts 2 at risk 3 filters
RAJLSKMKEN
Backlog 8
+
SES-412
Allow custom card types with user-defined icons
feature
5 pts
AI: readyJL
SES-418
Investigate slow query on pgvector cosine search
perfdb
3 pts
MK
SES-421
Write retro seeding prompts for Sonnet 4.5
docs
2 pts
RA
In Progress 5
SES-389
Agentic chat: tool-call streaming with partial renders
featureai
8 pts
3 sub-tasksEN
SES-401
Fix locked column override for admins
bug
2 pts
SK
SES-407
Burndown chart scale for multi-sprint view
feature
5 pts
JL
Review 3
SES-377 blocked by SES-401
Dependency enforcement on cross-list moves
feature
5 pts
RA
SES-383
Paystack webhook retry with exponential backoff
payments
3 pts
MK
Seshira AIsonnet · 14 tools
Plan sprint 23 — target 42 pts, balance workload across the team.
Agentic AI
Kanban · 5 views
Sprint planning
Anonymous retros
Gantt & timelines
OKRs & goals
Custom dashboards
Nested docs
GitHub PR auto-link
Automations
Card dependencies
Locked columns
SSO · SAML
White-label
Custom domains
Offline resilience
Custom roles
Audit logs
pgvector search
Real-time · WebSocket
Agentic AI
Kanban · 5 views
Sprint planning
Anonymous retros
Gantt & timelines
OKRs & goals
Custom dashboards
Nested docs
GitHub PR auto-link
Automations
Card dependencies
Locked columns
SSO · SAML
White-label
Custom domains
Offline resilience
Custom roles
Audit logs
pgvector search
Real-time · WebSocket
The AI, specifically

The AI doesn't just talk. It ships.

Most PM tools bolt on a chatbot. Seshira was built around one. Every capability below is actually in the product today.

14 TOOLS

Agentic chat that runs your workflow

Ask for a sprint — Seshira creates it. Ask for blockers — it finds and tags them. Every message can call real tools: create, move, assign, notify, search, escalate.

create_sprintmove_cardassign_membersdetect_blockersgenerate_pr_desc+9 more
Create a sprint for the billing v2 work and assign the team
Done. Here is what I did:
create_sprint"Billing v2" · Jul 1-14
assign_members4 members assigned
move_card11 cards added to sprint
DOC → CARDS

Turn any PRD into estimated tasks

Paste a PRD, meeting notes, or a bug report. Seshira reads it, clusters work into epics, writes cards, and estimates them using your team's actual velocity.

PRDs · meeting notes · plain textAuto-labeled (feat/bug/infra)Estimates from your velocity
SOURCE
PRD · billing-v2.md · 3,411 words
SES-412 gen
Stripe adapter: currency mapping
feat5 pts
SES-413 gen
Paystack webhook verification
infra3 pts
SES-414 gen
Unified event log schema
feat8 pts
SES-415 gen
Reconciliation cron
infra5 pts
SES-416 gen
Dunning emails
feat3 pts
SES-417 gen
Proration edge cases
bug2 pts
SPRINT INTEL

Velocity, blockers, burndown — predicted

Seshira watches every card and flags what's at risk before your standup. Predicted velocity, stale-card alerts, blocker radar, and a workload balancer that actually balances.

At-risk detectionVelocity predictionWorkload balancerStale-card alerts
BURNDOWN41/44 PTS
VELOCITYLAST 6 SPRINTS
Predicted next: 43 pts↑ 7%
AT-RISK CARDS2 detected by AI
SES-377Dependency enforcement on cross-list movesblocked 3dRA
SES-401Fix locked column override for adminsstale 5dSK
AUTOMATIONS

No-code triggers with AI actions

When cards enter Review, auto-generate PR descriptions from the diff. When a stale card sits 5 days, notify the owner. When a retro closes, draft the summary. Build it in English.

IF card enters Review → gen PRIF stale > 5d → notify ownerIF sprint closes → draft notes
IF
Card moves to Review
trigger · board.card.status_changed
AND
Card has linked GitHub branch
filter · card.github_branch ≠ null
Generate PR description from diff + card context
ai.generate · model=sonnet · credits ≈ 2
Post description to GitHub PR, notify reviewer
action · github.pr.update + notify
Everything you ship with

A full project OS. Not a toy.

BOARDS & VIEWS

5 views of the same reality.

Board, Table, Calendar, Timeline, Swimlane — pick the lens. Locked columns enforce RBAC. Dependencies block cross-list drags. Custom card types with your own icons and colors.

  • Kanban with WIP limits
  • Timeline & swimlane
  • Custom card types
  • Locked columns (RBAC)
  • Card dependencies with enforcement
Board/Sprint 22 · Kanban
Backlog3
SES-392
Custom card type: design review
feat
3 ptsEO
SES-395
Swimlane grouping by assignee
feat
5 ptsNW
SES-401
Fix locked column override for admins
bug
2 ptsSK
In Progress2
SES-377
Dependency enforcement on cross-list moves
feat
8 ptsRA
SES-384
Calendar view date-range picker
feat
5 ptsJL
Review2
SES-370
WIP limit warning toast
perf
2 ptsEO
SES-371
Board filter: label + assignee combo
feat
3 ptsNW
Done3
SES-360
Drag-and-drop reorder persistence
feat
5 ptsRA
SES-362
Inline card editing
feat
3 ptsSK
SPRINTS & RETROS

Sprint lifecycle, start to retro.

Planning, backlog refinement, velocity tracking, burndown charts. Then a 5-phase anonymous retro — write, reveal, vote, discuss, close — with AI seeding the first round.

  • Full sprint lifecycle
  • Velocity tracking
  • Burndown charts
  • 5-phase anonymous retros
  • Refinement assistant
Retro/Sprint 21 · Anonymous
1 · Write
2 · Reveal
3 · Vote
4 · Discuss
5 · Close
What went well
Agentic chat tools cut sprint planning to 12 min8
Paystack launch was smooth5
Burndown caught the slip early3
What to improve
Review column backs up every Thursday7
Too many P1 interrupts6
PRD-to-cards output needs more context2
Actions
WIP limit of 3 on Review
Interrupt budget = 20%
DOCS

Wiki-grade docs wired to cards.

Rich editor, nested pages, markdown, mentions, embeds. PRDs become cards with one command. Every doc links to the work it describes — and vice versa.

  • Nested pages
  • Markdown & slash commands
  • AI generates cards from doc
  • Mentions & embeds
  • Real-time collab
Docs/Engineering/Billing v2 · PRD
PRD · DRAFT · v0.4 · OWNED BY @ENYA

Billing v2: Paystack + Stripe under one gateway

We're unifying billing behind a single domain adapter so finance operations stay identical across regions. @morganspec'd the Stripe adapter in Q3; Paystack adapter needs parity on webhook semantics.

Goals

G1. Process subscriptions in NGN, ZAR, KES with zero reconciliation drift.
G2. Retry webhooks with exponential backoff up to 24h.
G3. Expose unified event log per tenant.

AI / Seshira · Generated 11 cards from this PRD
SES-412 · SES-413 · SES-414 · SES-415 · SES-416 · SES-417 · SES-418 · SES-419 · SES-420 · SES-421 · SES-422

Open questions

Do we want per-currency pricing displayed at checkout, or auto-convert at payment time?

PROJECTS & GOALS

Gantts, milestones, OKRs — connected.

Roadmap projects on a Gantt. Tie cards to milestones, milestones to OKRs, OKRs to company goals. One place to see what's slipping and why.

  • Gantt / timeline view
  • Milestones with dependencies
  • OKRs & goal hierarchy
  • Progress rolls up
  • Slip detection
Projects/Q2 roadmap · Timeline
W1
W2
W3
W4
W5
W6
W7
W8
W9
W10
W11
W12
Billing v2
6w
Agentic AI v3
8w
Mobile App
4w
Whitelabel
4w
SSO / SAML
4w
DASHBOARDS

Drag-and-drop widgets. Real metrics.

Build a dashboard per team, per project, per quarter. Velocity, cycle time, throughput, AI credit burn. No ETL — it's just your data.

  • Custom widgets
  • Drag & drop layout
  • Velocity, cycle, throughput
  • AI credit metering
  • Per-team views
Dashboards/Team pulse
Velocity · 6w avg
38.2
↑ 14% vs prev
Cycle time
3.4d
↓ 0.8d
AI credits
2,147
of 5,000 / mo
Throughput · cards done / day
Mon · Jan 13Sat · Jan 25
Blockers
2
auto-flagged
BUILT FOR TEAMS

Enterprise guardrails without the enterprise friction.

SSO/SAML, custom roles with granular permissions, white-label branding, custom domains, 2FA, guest access, audit logs. Offline changes queue locally and auto-sync on reconnect.

  • SSO / SAML
  • Custom roles (granular)
  • White-label (logo, AI name, hue)
  • Custom domains
  • Audit logs · 2FA · guest access
  • Offline resilience
Automations/PR description generator
IF
Card moves to Review
trigger · board.card.status_changed
AND
Card has linked GitHub branch
filter · card.github_branch ≠ null
Generate PR description from diff + card context
ai.generate · model=sonnet · credits ≈ 2
Post description to GitHub PR, notify reviewer
action · github.pr.update + notify
ran 14× this weekavg 1.8s0 errors
Pricing

Pricing right on the landing page.

Because you shouldn't have to click "contact us" to see a number.

Free
$0forever

For small teams trying the agentic workflow.

Start free
  • Up to 10 members
  • 5 boards
  • 100 AI credits / month
  • All 5 board views
  • Sprints, docs, retros
  • Community support
RECOMMENDED
Pro
$12per user / month

Unlimited everything. Made for growing teams.

Start 14-day trial
  • Unlimited members & boards
  • 5,000 AI credits / month
  • Custom roles & permissions
  • GitHub PR auto-linking
  • Automations (no-code AI triggers)
  • Custom dashboards
  • Priority support
Business
$25per user / month

White-label, SSO, and the big-org checklist.

Talk to sales
  • Everything in Pro
  • Unlimited AI credits
  • SSO / SAML (Okta, Azure AD, Google)
  • White-label branding
  • Custom domains
  • Audit logs · 2FA · guest access
  • Dedicated success manager
  • 99.95% SLA
Billing in USD via Stripe · in NGN · ZAR · KES via Paystack.
How it works

Three steps. No onboarding call required.

01

Import or start fresh

CSV from Jira, JSON from Linear, or build a board from scratch in 30 seconds. Every import preserves labels, estimates, and history.

02

Let the agent watch

Seshira reads your backlog, sprints, and docs. Within a day it knows your velocity, your stale-card patterns, and who usually owns what.

03

Ship with 14 tools

Ask for sprints, retros, PR descriptions, or blocker reports. The agent does it. You review and merge.

Built on

Boring infrastructure, interesting ideas.

Next.js 16
React 19
TypeScript
Tailwind
Go · chi
PostgreSQL
pgvector
AWS Bedrock
Claude Sonnet
Claude Haiku
Titan Embeddings
Redis
WebSocket
Stripe
Paystack
FAQ

Common questions.

The AI has access to 14 real tools — create_sprint, move_card, assign_members, detect_blockers, generate_pr_description, and more. It doesn't just suggest; it executes (and shows you what it did so you can undo).

Yes. We support CSV, JSON, and native API imports for all four. Labels, estimates, comments, and status history come across intact.

Claude Sonnet for complex reasoning (sprint planning, PRD-to-tasks), Claude Haiku for fast operations (search, summarization). Titan Embeddings for semantic search across your workspace.

One agent action that hits a tool, or roughly one Sonnet completion. Free gets 100/mo, Pro 5,000/mo, Business unlimited. Most teams use ~1,000–2,000/mo.

No. We use AWS Bedrock, which contractually prohibits training on customer data. Your workspace stays yours.

Changes queue locally in IndexedDB and auto-sync on reconnect. Cards edited offline are marked until they're merged, and conflicts surface as a review step.

Not today. Business customers get their own isolated tenant with optional EU data residency. Self-host is on the 2027 roadmap.

Yes. On Business, rename "Seshira AI" to whatever fits your brand, swap the accent color, add your logo. Your users see your product.

Branch naming convention maps to card IDs (e.g. ses-412-stripe-adapter). PRs auto-link to the card, status updates flow both ways, and automations can generate PR descriptions.

Stripe for global cards. Paystack for NGN, ZAR, KES, GHS with native bank transfer and mobile money.

AI features pause until the next billing cycle, or you can top up in 1,000-credit packs. Non-AI features keep working normally.

Yes — REST + WebSocket for real-time events. Webhook signatures, OAuth apps, and per-workspace rate limits. Docs at seshira.dev.

SSO/SAML on Business with Okta, Azure AD, Google, OneLogin. SCIM provisioning included.

Morphlix — a small team of engineers who got tired of PM tools that pretend to be smart.

Your team's AI-powered
engineering OS.

10 minutes to set up. 14 tools that execute. Every plan free to try.

NO CREDIT CARD · 10 USERS FREE · 100 AI CREDITS / MO