The 2026 Cap Table & Portfolio Operations Playbook
Seven workflows where Finpy Asset Manager handles the cap table mechanics, deal pipeline, fund admin, and LP-side view on a single data model. Cross-org entities, double-entry security transactions, on-the-fly IRR and MOIC, audit ledger by default, and a Model Context Protocol layer that no other platform ships.
Finpy Team · June 22, 2026
Key takeaways
- 1.The wedge. Cap table mechanics, deal pipeline, fund admin, and LP view sit on the same entity graph. No bolted-on integrations between four tools.
- 2.Dual perspective. The cap table view (issuer side) and the position view (stakeholder side) read the same rows from opposite ends. Cross-org access is first-class.
- 3.Per-entity pricing. Your first entity is free. Each additional entity is $6 per month. No per-seat charge for stakeholders or LPs.
- 4.MCP-native. Connect Claude Desktop to mcp.finpy.tech/mcp and query your cap table in conversation. No competitor ships an MCP server.
Competitive map
Six tools cover slices of the workflow well. Each breaks somewhere important. The final row is what Finpy Asset Manager does about it.
Why a fund operations team is the one writing this
Every workflow below ties back to a single data model. Cross-org entities, double-entry security transactions, on-the-fly performance computation. That is the engine we run at Finpy Asset Manager, the platform behind this analysis. Open the live dashboard and the same workflows are one click away.
Open the live dashboardThree workflow threads
Seven case studies grouped by what they exercise. Fundraising mechanics, portfolio operations, and an AI-native workflow that no competitor matches.
Fundraising mechanics
Cross-org Seed raises, multi-tranche Series B deals, and syndicate workflows. The data model treats the cap table as a double-entry ledger and lets stakeholders own their own accounts.
Portfolio operations
Fund-level performance computation across many portfolio companies. Look-through ownership for fund-of-funds. Structured financial data per portco rolling up to fund-level NAV.
AI-native workflows
Agentic AI access to fund data through the Model Context Protocol. The cap table, the deal pipeline, the income statements all become first-class context for Claude, Cursor, and any other MCP-aware client.
What Finpy Asset Manager delivers
Dual-perspective cap table. Every entity carries both the issuer view (this is my cap table) and the stakeholder view (these are my positions). Cross-org access is a first-class concept, not a workaround. A founder, a fund, and an LP all see the same transaction from their own angle.
Double-entry security ledger. Every transaction records the debit side and the credit side, in both units and dollars. The cap table is reconstructed by adding up the transactions, not by snapshotting and diffing. Partial transfers, complex conversions, and warrant exercises all express cleanly.
Per-entity pricing. First entity free; each additional entity $6 per month. No per-stakeholder surcharge. A fund running 20 portfolio companies pays $114 per month including the fund entity itself. Pricing does not penalize fundraising.
The cross-org Seed raise
A startup raises Seed without paying a per-stakeholder surcharge for each investor on the cap table.
Each stakeholder on the cap table points back to a separate Finpy account owned by the investor. The issuer pays nothing extra for adding stakeholders. Every LP signs up on their own free entity and sees their positions from their side. The same cap table row renders differently depending on who is reading it.
Finpy
- Per-entity pricing leaves stakeholder count unconstrained
- LPs maintain their own accounts at no cost to the issuer
- Dual-perspective view: issuer reads the cap table, LP reads positions, both reference the same rows
- Cross-org access controlled per role: Owner, Admin, Editor, Viewer
Typical competitor
- Carta typically charges per stakeholder for active cap tables
- Per-seat models punish founders for fundraising success
- Closed platforms make LPs subordinate accounts of the issuer
The pricing wedge is real and the dual-perspective model is the structural reason it works.
Series B with five tranches
One funding round issues five distinct securities, each with its own terms. The cap table reconciles by double-entry, not snapshot diff.
Asset Manager handles every security type on one unified schema. A multi-tranche Series B issues five distinct securities (Series B-1 Preferred at $20 with 1x liquidation, Series B-2 with 1.5x participation, Series B-3 convertible, and two warrant tranches). Each security gets its own transactions with debit and credit recorded in both units and dollars. The cap table is the running total of these transactions at a point in time.
Finpy
- Seven security types on one schema: common, preferred, convertible, SAFE, warrant, option, bond
- Every transaction records both sides of the trade in units and in dollars
- Securities tie back to the funding round that issued them; deleting the round is blocked while transactions exist
- Cap table view computed on demand from transactions, not stored as a snapshot
Typical competitor
- Most cap table tools collapse a multi-tranche round into one security row with notes
- Snapshot-based platforms cannot express partial transfers or staged conversions cleanly
- Single-security-per-round assumption fails on real venture rounds
The double-entry ledger is the reason cap table mechanics survive contact with messy round structures.
The syndicate with cross-org members
A syndicate vehicle has members from five different organizations. Each member sees the syndicate positions; the carry waterfall splits on the stakeholder distribution tiers configured at setup.
A syndicate is just an entity set up as a fund. Its members reference outside Finpy accounts; each member retains full ownership of their own organization. The carry waterfall is built in: ordered distribution tiers, carried interest percentages, and catch-up rates are all native fields on each stakeholder. Members can leave; the syndicate continues; the data does not move.
Finpy
- Members bring their own accounts; the syndicate cannot lock them in
- Carry waterfall is native: ordered distribution tiers split distributions automatically
- Cross-org membership controls access without merging organizations
- Cap table view per member shows only what their role permits
Typical competitor
- AngelList Stack runs a closed roll-up vehicle ecosystem
- Closed platforms make member exit difficult and data export limited
- Outside syndicate members cannot bring their existing accounts
Open data ownership is the difference between a platform and a closed network.
Fund with 20 portfolio companies, IRR computed live
Each portfolio company is a tracked Entity. Performance metrics (IRR, MOIC, TVPI, DPI) compute on demand from the same data the cap table runs on.
Each portfolio company is a holding linked back to a target entity and a parent fund. Investments and distributions record with full debit and credit semantics. IRR, MOIC, TVPI, and DPI compute at the moment of the query, not in a nightly batch. A new mark-to-market triggers fresh performance immediately.
Finpy
- Performance computation runs on the same data the cap table reads
- No overnight batch; the dashboard reflects the latest fair value as soon as it lands
- Each cash flow carries a flag for whether it belongs in the IRR calculation
- Optional point-in-time performance snapshots when needed for audit
Typical competitor
- Enterprise fund admin platforms typically batch IRR overnight
- Reconciliation between cap table system and fund admin system is the standard pain
- Portfolio companies as PDF uploads rather than structured entities
The same data drives the cap table and the IRR. The reconciliation problem dissolves.
Fund-of-funds look-through
Fund A holds Fund B; Fund B holds four portcos. The look-through query traverses both levels and surfaces underlying exposure to a specific company.
Fund A is a stakeholder on Fund B cap table. Fund A also has a holding pointing at Fund B as a position. Fund B in turn holds four portfolio companies. The look-through query walks Fund A holdings, identifies the ones whose target is itself a fund, and resolves to the underlying portco set with computed ownership percentages.
Finpy
- Entities can reference each other as parents, modeling multi-level fund structures natively
- A holding link preserves the relationship even if the sub-fund later leaves the platform
- Look-through is a property of the data model, not a bolted-on report
- Family offices and fund-of-funds get the same traversal as primary funds
Typical competitor
- Single-entity cap table tools cannot represent multi-level fund structures
- Investor-relations platforms treat sub-fund relationships as PDF attachments
- Reporting tools that pull from accounting systems lose the entity graph
Look-through ownership only works if the data model is a graph. Most competitors model it as a list.
Portfolio financial drill-down
Each portfolio company files an income statement, a balance sheet, and a cash flow statement every quarter. Custom KPIs (ARR growth, gross margin, burn) roll up to fund-level valuations.
Financials enter as structured rows per portfolio company per period: income statement, balance sheet, cash flow statement, and a derived metrics layer. KPIs are defined once and tracked quarter over quarter. Updated valuations attach to the holding; the fund-level NAV reads the aggregated valuations across the portfolio.
Finpy
- Four financial statement types fully structured (income, balance, cash flow, metrics)
- Custom KPIs let each portfolio company report what matters in its business model
- Each KPI value carries a scenario tag: actual, forecast, or budget
- Updated valuations feed both the cap table and the fund NAV
Typical competitor
- Most fund admin platforms accept financials as PDF uploads
- PDF financials cannot be queried, compared, or aggregated
- Portfolio reporting tools treat KPIs as static dashboard widgets
Structured rows beat PDFs. The fund admin platforms that treat financials as documents lose the analytics layer.
Show me my cap table (Claude + MCP)
A user opens Claude Desktop, connects mcp.finpy.tech/mcp, and types "show me my Series B cap table". The agent calls the right tool and renders the result.
The MCP server is a first-class surface on the same data the dashboard reads. It exposes 67 tools covering create, read, and update across every Finpy Asset Manager entity. The user types a natural question; Claude selects the tool; the answer is the structured data, not an opinion. The same server handles "create a stakeholder John Doe", "set the fair value of Acme Corp to $10M", and "show me Q4 revenue across all portcos".
Finpy
- OAuth device flow for clean authentication from Claude / Cursor / Code
- 67+ tools cover the full CRUD surface across cap table, deals, holdings, financials
- Same data model as the dashboard; no separate AI export pipeline
- Tool selection driven by the agent; the user does not memorize API endpoints
Typical competitor
- No competitor ships an MCP server today
- Carta has an API but no agentic interface
- Most fund admin platforms only expose data via PDF reports or CSV exports
The MCP wedge will not be a wedge forever, but as of June 2026 no other platform ships it.
What links these seven
Seven workflows. One data model. Same Entity graph underneath every case.
- ·Entity graph. Funds, companies, individuals, and syndicates all live on one shared schema, with parent-child relationships built in. Fund-of-funds and family office structures are first-class, not edge cases.
- ·Double-entry ledger. Security transactions and portfolio cash flows share the same debit and credit convention. The cap table and the fund performance read the same data.
- ·Cross-org access. Memberships and invitations span organizations natively. Stakeholders own their own accounts as a structural property of the platform, not as a workaround.
- ·Audit by default. Every record carries timestamps and tracks who created, updated, or deleted it. The audit log captures before-and-after snapshots of every change. The compliance team does not need a separate tool.
Try it yourself
Open the live dashboard, create your first entity (free), and run any of the seven workflows above. For the AI-native one, connect Claude Desktop to the Finpy MCP server and type the question you would ask your fund analyst.
Open the live AssetManager dashboardThe MCP server is at mcp.finpy.tech/mcp. Connect from Claude Desktop, Cursor, or any MCP-aware client. The companion Factor Markets engine is on the 2026 Value Landscape article.
How the data model works
Every record on the platform carries the same five things: a created timestamp, an updated timestamp, an optional deletion timestamp, plus the user identifiers of whoever created, updated, or deleted it. Deletions are soft by default; the application layer never hard-deletes. Ownership relationships cascade when the parent goes away; financially critical relationships (funding rounds, securities) refuse to delete while transactions depend on them; audit-preserving relationships keep the record alive even after the target it pointed to leaves the platform.
The cap table is reconstructed by adding up the security transactions for an entity at a point in time. The fund performance is computed by aggregating the cash flows for a fund through a chosen report date. Optional point-in-time snapshots exist for audit and reporting; the live view does not depend on them. KPI values store quarterly metrics with full scenario support (actual, forecast, budget) so the same number can carry three readings. The MCP server is a thin layer on the same backend endpoints the dashboard uses; no separate AI pipeline.
What could go wrong
- ·Linear pricing. $6 per entity per month adds up. A fund with 100 portfolio companies pays around $600 monthly. That is still well below enterprise fund admin pricing, but the per-entity model is not free.
- ·Not a substitute for an audit firm. The audit ledger helps reconcile and trace changes; it does not produce a regulator-signed opinion. Year-end audit still happens with a CPA firm.
- ·MCP integration is new. The protocol itself is months old. Tooling matures monthly; some clients have rough edges. The dashboard remains the primary surface for users who prefer a UI.
Built by a team that runs the platform every day
Finpy Asset Manager unifies cap table mechanics, deal pipeline, fund admin, and the LP-side view on one entity graph. Per-entity pricing, no per-seat ceiling. First entity free. Built for emerging managers and multi-fund family offices, opened to anyone who wants the same workflow without the legacy fund admin price tag.
Explore Finpy Asset ManagerAuthor positioning
Finpy Team. Finpy Asset Manager is a paid product from Finpy Tech. The dashboard view shown in the examples is the same view paying customers use. Pricing is per-entity (first entity free, $6 per additional entity per month). Competitor claims sourced from public product pages and review sites as of June 22, 2026.
Nothing on this page is legal, tax, or investment advice. The data model details described above are accurate to the platform as of the publish date and will evolve. The MCP integration is in active development. Workflow descriptions are not promises of future features. Talk to your own fund admin, tax, and legal advisors before relying on any platform for fund operations.