Kaynak doğrulamalı araştırma: Jupiter / Raydium

Verified source facts

1. Jupiter describes Swap V2 /order plus /execute as its recommended Meta-Aggregator flow, with Metis, JupiterZ, Dflow and OKX routing engines competing for an order. 2. The V2 /order response includes an unsigned transaction, requestId, expected outAmount and winning router; /execute reports success or failure and actual input and output amounts. 3. Jupiter says the older Metis Swap V1 quote/swap API is no longer actively maintained and has been superseded by Swap V2. Its legacy fields remain useful only when the operator deliberately uses that older flow. 4. Raydium documents atomic single-hop, multi-hop and split routing across its pools. 5. Raydium requires integrators to handle stale quotes, per-hop minimum amounts and transaction failure instead of treating a displayed quote as guaranteed execution.

API version boundary

Do not mix fields from Jupiter's legacy V1 quote response with the recommended V2 execution contract. For a current integration, record whether the request used V2 /order and /execute or V2 /build, which router won, whether optional parameters restricted routing, the signed minimum output, and the actual output returned after execution.

Comparison at a glance

1. Compare token identity on the same job

A platform comparison becomes useful only when both venues are measured against one job. For token identity, that job concerns input and output mint addresses. Marketing pages frequently describe broad capability while the actual decision lives in a specific pair, contract, network, entity, or account tier. Build two evidence rows from the same timestamp, then verify both official mint addresses. Until that check is complete, neither venue earns the point.

The comparison should distinguish a published ceiling from an executable outcome. A fee table does not reveal slippage; a product menu does not prove regional eligibility; a security feature page does not prove the control is enabled on the reader's account. Record the route from public statement to authenticated evidence. The unresolved variable is spoofed-token listings, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For token identity, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how input and output mint addresses changes the entry, exit, monitoring, or fallback decision, and make the operator verify both official mint addresses. The remaining uncertainty—spoofed-token listings—belongs in the dated scorecard. That produces a job-specific preference instead of a permanent winner.

2. Compare liquidity scope on the same job

For liquidity scope, that job concerns DEX sources and pools searched. Build two evidence rows from the same timestamp, then compare the actual route plan.

The unresolved variable is future pool availability, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For liquidity scope, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how DEX sources and pools searched changes the entry, exit, monitoring, or fallback decision, and make the operator compare the actual route plan. The remaining uncertainty—future pool availability—belongs in the dated scorecard.

3. Compare quote mode on the same job

For quote mode, that job concerns ExactIn versus ExactOut. Build two evidence rows from the same timestamp, then match the mode to the user's job.

The unresolved variable is AMM-specific support, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For quote mode, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how ExactIn versus ExactOut changes the entry, exit, monitoring, or fallback decision, and make the operator match the mode to the user's job. The remaining uncertainty—AMM-specific support—belongs in the dated scorecard.

4. Compare minimum output on the same job

For minimum output, that job concerns slippage threshold and failure boundary. Build two evidence rows from the same timestamp, then record the signed minimum.

The unresolved variable is price movement before landing, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For minimum output, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how slippage threshold and failure boundary changes the entry, exit, monitoring, or fallback decision, and make the operator record the signed minimum. The remaining uncertainty—price movement before landing—belongs in the dated scorecard.

5. Compare price impact on the same job

For price impact, that job concerns size relative to executable depth. Build two evidence rows from the same timestamp, then re-quote at the intended size.

The unresolved variable is stress-period liquidity, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For price impact, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how size relative to executable depth changes the entry, exit, monitoring, or fallback decision, and make the operator re-quote at the intended size. The remaining uncertainty—stress-period liquidity—belongs in the dated scorecard.

6. Compare route complexity on the same job

For route complexity, that job concerns single, multi-hop or split execution. Build two evidence rows from the same timestamp, then inspect every intermediate asset.

The unresolved variable is intermediate pool failure, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For route complexity, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how single, multi-hop or split execution changes the entry, exit, monitoring, or fallback decision, and make the operator inspect every intermediate asset. The remaining uncertainty—intermediate pool failure—belongs in the dated scorecard.

7. Compare priority fee on the same job

For priority fee, that job concerns transaction landing cost and cap. Build two evidence rows from the same timestamp, then set a bounded fee policy.

The unresolved variable is future congestion, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For priority fee, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how transaction landing cost and cap changes the entry, exit, monitoring, or fallback decision, and make the operator set a bounded fee policy. The remaining uncertainty—future congestion—belongs in the dated scorecard.

8. Compare quote freshness on the same job

For quote freshness, that job concerns slot and time between quote and signing. Build two evidence rows from the same timestamp, then refresh before approval.

The unresolved variable is wallet delay, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For quote freshness, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how slot and time between quote and signing changes the entry, exit, monitoring, or fallback decision, and make the operator refresh before approval. The remaining uncertainty—wallet delay—belongs in the dated scorecard.

9. Compare simulation on the same job

For simulation, that job concerns transaction accounts and compute. Build two evidence rows from the same timestamp, then simulate before broadcast.

The unresolved variable is runtime state changes, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For simulation, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how transaction accounts and compute changes the entry, exit, monitoring, or fallback decision, and make the operator simulate before broadcast. The remaining uncertainty—runtime state changes—belongs in the dated scorecard.

10. Compare post-trade proof on the same job

For post-trade proof, that job concerns signature, balance delta and explorer status. Build two evidence rows from the same timestamp, then reconcile expected and received output.

The unresolved variable is downstream token risk, so the scorecard must preserve that blank instead of awarding a speculative advantage.

For post-trade proof, connect the finding to the reader's actual order plan rather than the other nine rows in the abstract. Document how signature, balance delta and explorer status changes the entry, exit, monitoring, or fallback decision, and make the operator reconcile expected and received output. The remaining uncertainty—downstream token risk—belongs in the dated scorecard.

Decision scenarios

Scenario 1: when token identity becomes decisive

Consider a reader whose main constraint is input and output mint addresses. The correct response is not to infer a universal platform or product winner. It is to verify both official mint addresses, retain the evidence, and define a stop condition before funds or permissions are committed. If the evidence changes, the decision must be reopened. This produces a repeatable operating rule rather than a one-time opinion.

Scenario 2: when liquidity scope becomes decisive

Consider a reader whose main constraint is DEX sources and pools searched. It is to compare the actual route plan, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 3: when quote mode becomes decisive

Consider a reader whose main constraint is ExactIn versus ExactOut. It is to match the mode to the user's job, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 4: when minimum output becomes decisive

Consider a reader whose main constraint is slippage threshold and failure boundary. It is to record the signed minimum, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 5: when price impact becomes decisive

Consider a reader whose main constraint is size relative to executable depth. It is to re-quote at the intended size, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 6: when route complexity becomes decisive

Consider a reader whose main constraint is single, multi-hop or split execution. It is to inspect every intermediate asset, retain the evidence, and define a stop condition before funds or permissions are committed.

Sources and editorial method

Product availability, fees, controls, and legal terms can vary by entity, region, account tier, and time. Verify the live official page and the exact account screen before acting.